Artificial Intelligence for Clean Industry: A Multi-Agent Framework for Sustainable Waste, Energy, and Emission Practices in Gazipur Factories

Artificial Intelligence for Clean Industry: A Multi-Agent Framework for Sustainable Waste, Energy, and Emission Practices in Gazipur Factories | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Artificial Intelligence for Clean Industry: A Multi-Agent Framework for Sustainable Waste, Energy, and Emission Practices in Gazipur Factories Md. Arifur Rahman, Efat Ara Haque This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-7278542/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The paper examines the ways of how artificial intelligence can be utilized in maximizing the sustainability of industry and municipal waste treatment systems. Applying the case of automotive assembly line operations and municipal waste management in Gazipur City Corporation of Bangladesh, we show the potential of AI in resolving resource optimization issues on various levels. In analysis done on automotive assembly line, Genetic Algorithm (GA) optimization was used to reduce energy consumption and waste generation, whereas the municipal waste management analysis showed that there are huge optimization chances in certain cities, such as Gazipur where 1747.121 tons of municipal wastes generate each day with just 60 percent collection efficiency. In an industrial context, GA optimization resulted in a 10 % energy decrease (specifically, ~2200 kWh to ~2000 kWh per run) as well as a 25 percent decrease in waste (particularly, ~22 units to ~17 units). These findings show that AI has the potential to help in the ineffective processes in manufacturing as well as managing municipal solid waste. The results indicate that analogous optimization plans ought to enhance waste collection performance in a fast-growing city such as Gazipur City Corporation, which is projected to experience an increase in waste generation up to 5496.80 tons/day by 2035. The findings of this research can serve as evidence-based information that can be offered to the decision-makers to implement AI-based sustainability solutions in industrial and municipal settings. Environmental Engineering Artificial intelligence genetic algorithm sustainable manufacturing municipal solid waste environmental impact machine learning Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Figure 6 Figure 7 Figure 8 Figure 9 Figure 10 1 | INTRODUCTION Sustainability of industrial operations and management of municipal wastes are some of the most important issues directly facing developing countries because of the increased rate of urbanization and industrialization that put unprecedented pressures on the environment (Ragai et al., 2022). Over 70% of the chemical waste produced by automotive factories is recognized as hazardous. Most of the waste weight comes from scrap metal, while paint solvents and other trash help fill in the rest (Pei et al., 2021; Tao et al., 2021). While many things are recycled, reports note that problems remain. There is less pollution, a drop in chemical waste, and fewer unproductive processes(Greif et al., 2020). In place of this, many companies now aim to make their operations more sustainable, which has major effects on automobile manufacturing. A lot is being done today to lower energy and scrap production. By identifying and repairing problems and applying the circular economy method to use available parts and materials(Jadhav & Mantha, 2021; Khan, 2019). The heavy use of energy and production of wastes during the automotive industry represent a wider context of the problems of sustainability of urban centers as they continue to expand. In Bangladesh, municipal solid waste management has turned out to be very tough with cities like Gazipur City Corporation producing about 1747.121 tons/day of solid waste with the collection rate of the organization standing at 60 percent. Gazipur City Corporation, which is the biggest city corporation in Bangladesh, is an example of the two-fold challenge of industrial and municipal sustainability issues1. Being one of the key industrial centers with population of 3,979,775, the city has to deal with both a manufacturing and a household pressure related to water pollution(Pratticò & Lamberti, 2021). The composition of the waste in the city displays a composition of 90.54 percent organic content having a moisture content of 49.04 percent posing special problems in processing and disposal1. The projection of population of 12,521,190 by the year 2035 is likely to cause waste generation to be increased to 5496.80 tons per day ( CEE_105426,09,49.Pdf , n.d.). Manufacturing industry, especially the car manufacturing sector, is a heavy energy consumer and emitter of major waste flows. Computers and motors consume almost the same proportion of energy used in factories and more than 70 percent of the automotive. Recycling is confronted with the same issue as municipal waste management, in that proper collection systems, processing, and disposal represent a risk to the environment and health of civilians. Smart manufacturing technologies and artificial intelligence have potential as solutions to the problems of sustainability in industries as well as municipalities(Xuan Ao et al., 2022). AI permits vast accuracy in resource planning, power management during low activity times, and automations of surveillance systems. Similar applications could be made in a municipal situation to optimize waste collection routes, bulk generation trends, and enhance efficiency of processing. This research explores AI optimization methods that can be applied to both the automotive manufacturing industries, as well as, to municipal waste management practices, and Gazipur City Corporation serves to exemplify how the idea of urban sustainability applies in developing nations. 1.1 | REVIEW OF LITERATURE 1.1.1 | Sustainable Production in the Automotive Field Many areas of manufacturing are considered environmentally friendly for making production greener. This happens in the automotive sector just as it does in the aerospace sector. Saving material and energy at all stages of a vehicle’s life (Ozturk & Dincer, 2021; Singh & Basheer, 2023). The rise of these impressive figures in metal recycling has not reduced emissions. Studying the issue is the task of researchers at the EPA. Industry mitigation measures do not decrease the amount of waste produced, which is often very high. The paint-related volatile organic compounds are now the greatest contributor to all other types of air emissions. Because of this, manufacturing vehicles using less energy and reducing waste is still a priority(Wan et al., 2021; W. Zhang et al., 2021). Various researchers suggest that companies would benefit from using lean manufacturing, the circular economy, and digital systems at the same time. Approaches that apply new technology can help deal with these issues. Using simulation and AI plays a key role in reaching environmental targets in complicated manufacturing systems(Han & Zhao, 2022; Khosravani et al., 2023). 1.1.2 | Use of AI and data for sustainability AI and data are being used to make the world more sustainable. Many experts believe that AI/ML can improve how sustainable product development is done today. Most say that AI and ML can cut down on the energy used and the waste produced in manufacturing, an expert noted (Fang et al., 2020; Weyand et al., 2021). AI can optimize the way we look after energy resources. It monitors changes and enables to adjust use as occur (H. Zhang & Liu, 2022). Experts have discovered that using artificial intelligence helps the industry grow. Automobile energy methods help lessen emissions, but still do as well as previous methods for chemical processing(Asadabadi et al., 2020). Organizations in smart manufacturing environments are also exploring the employment of AI in helping them to streamline supply chains, hence reducing wastage and enhanced utilization of resources(A. Kumar & Singh, 2021; Shuvo & Manogharan, 2021). 1.1.3 | Methods for Simulation and Optimization People commonly use simulation models to test how well-proposed sustainability actions will function. In manufacturing, this type of engineering is used a great deal in research and development. Numerous experts believe that simulation modeling helps assess and improve sustainability when it comes to making and transporting products(Bany Salameh et al., 2020; Eunike et al., 2022). It serves to identify and understand future developments. Experiments that do not disrupt what is going on in the network. Via discrete and continuous approaches, energy flow, material flow, and feedback loops have been modeled using techniques from event-driven, agent-based, and system dynamics methods. Production processes cause changes in the environment. (Fathollahi-Fard et al., 2021; Oke, 2023). It generally makes it convenient for people to test their ideas. Genetic algorithms are regarded by many users as important AI tools. Efforts to continuously improve how things are done at work. Recently, literature in the field often deals with genetic algorithms, particle swarms, and reinforcement learning. Being able to arrange worker shifts to use the most energy and produce the least waste (Chen et al., 2023; Chesani et al., 2020). For example, the combination of GA with Bayesian optimization can find nearly optimal solutions in a way that saves the most energy but does not drop quality(Roy & Das, 2022; Xuan Ao et al., 2022; W. Zhang et al., 2021). 1.1.4 | Using AI in the automotive manufacturing industry Among different industries, digital twin and smart factories are of the greatest importance in automotive manufacturing. Concepts are proving to be very important. Because these lines can be virtual, significant parts of the production process can be analyzed—without changing the actual line—to find better ways to manage energy and waste( Application of Artificial Intelligence in Sustainable Manufacturing , n.d.; Vijay Kumar & Shahin, 2025). Data contained in reports indicates that brands in the auto industry are focusing on using AI to cut energy consumption. Material waste that happens at the time of use( Application of Artificial Intelligence in Sustainable Manufacturing , n.d.; Vijay Kumar & Shahin, 2025). In addition, practical methods have found that AI can help raise equipment effectiveness (OEE) by several percent in the automotive sector(Vijay Kumar & Shahin, 2025). Neural networks are often used in different kinds of work. Many studies focus on models tailored for industries, yet there were few that applied known AI approaches such as symbolic or search-based AI, in supporting sustainability for manufacturing. Because of these requirements, we choose GA for our genetic algorithm. After that, the idea of optimization is discussed, along with how a manufacturing simulator is used. 1.1.5 | Development Countries Municipal Solid Waste Management The management of municipal solid waste is a very critical sustainability issue in fast developing–urbanizing countries. The case of cities such as Gazipur City Corporation points at the complexity of urban waste management in manufacturing cities where the both production activities and the residential population increase the problems of managing waste. Most countries in the developing world do not have the infrastructure or the technological potential of managing the ever-rising amounts of waste using the current waste management practices. In Gazipur City Corporation there are five zones within the geographical area, i.e. Gazipur, Kaligonj, Kaliakoir, Kapasiya and Sreepur, that are covered under the waste management system, yet only a 60 percent efficiency marks successful collection. This ineffectiveness leads into indiscriminate litter in open environment, roadside and drainage systems causing environmental degradation and health hazards. Municipal solid waste in cities such as Gazipur has significantly different composition in comparison to the industrialized nations, that is; high organic composition (85–95 percentage) and moisture content (45–55 percentage)( CEE_105426,09,49.Pdf , n.d.). The composition poses special problems in terms of processing and disposal, and it needs dissimilar technological procedures when compared to the developed nations1. Its organic content is high, and it has an opportunity of composting and production of biogas, and is dependent on transport and processing performances due to moisture content. Most of these challenges could be mitigated using AI applications in municipal waste management such as route optimization, forecasts of demand, and efficiency in the processing of garbage. This issue of scalability of AI-related solutions makes it especially applicable to fast-growing cities such as Gazipur, where the usual means of waste management are unable to withhold the process of urbanization( CEE_105426,09,49.Pdf , n.d.). 2 | METHODOLOGY In order to illustrate the wider value of AI optimization strategies, we employed our methodology to include the workaround management of municipal waste, with Gazipur City Corporation serving as a reference point study area. The process of separating municipal wastes can be modeled same way, with the collection zones corresponding to the manufacturing stations. Gazipur City Corporation is serving its population in five major zones Gazipur, Kaligonj, Kaliakoir, Kapasiya and Sreepur. Different waste generation characteristics, as well as collection efficiency rates and processing abilities apply to each zone. Similar GA methods may be utilized to optimize the system with the following goals: Reduction of distances of collection routes and fuel consumption Optimizing the collection (which is 60 percent at present) Optimization of the use of processing capacity Environmental positive impact by means of better disposal means The rate of waste generation per capita, which is 0.439 kg per capita per day, and the total daily ratio; 1747.121 tons, offer base value considerations on which optimization can take place1. Such future projections as 5496.80 tons per day in 2035 provide high concerns with the need to implement efficient management systems. We made a simulation model for a simplified automotive assembly line, using discrete events. The model consists of four stations, each representing important steps (stamping, welding, painting, assembly). Every station includes a tunable speed factor p i that alters how fast it operates, (relative to the default setting). We expect that going faster will allow more products to be processed. Yet, it also leads to more power and a larger number of manufacturing flaws. In particular, at station i , the scrap probability increases as the velocity p i goes up and more waste is found. Energy use per part is illustrated as E i = E base, i × p i , to show that quicker outputs require more energy. We chose values for E base, i , and scrap baselines to represent average plant behavior (stamping is estimated to use 5 kWh/unit, painting 8 kWh/unit) (Han & Zhao, 2022 ; Khosravani et al., 2023 ; P. Kumar, 2023 ; Xia et al., 2023 ). A simulation is performed for N = 100 vehicles in the production run. With each vehicle and each station, the model adds energy cost E i and, with probability s i , adds that vehicle to a group of defective parts. As a result, the model provides outcomes for total energy consumption E total and defective count W. Our goal is to minimize both of these numbers, reflecting less waste(Ragai et al., 2022 ). The optimization of AI within our system takes place thanks to the use of a genetic algorithm (GA). Solutions are recorded in vectors p = [p1, p2, p3, p4 ] and the range for the speed is 1.0 to 3.0. A group of 50 individuals is worked with by the GA across over 100 generations. We can determine fitness by using F = E total + 100 ×W and it costs much more if the facility generates more waste. With every generation, selection, crossover, and mutation happen and elitism maintains the best results. The basic principles behind this GA are not neural networks, gradient descent, or certain regression—it relies on general optimization instead. That’s why we favor using a general AI approach in our work(Xia et al., 2023 ). Early tests allowed us to choose population size and mutation rate settings that would help the population converge quickly. To suppress simulation noise, each candidate score is made by taking the average of various simulation results. We repeat some experiments using GA to make sure our approach is reliable. We assume that no AI optimization is the main type of case. The angular speed is chosen as p i =1.0. The optimized GA results for energy and waste are compared to the previous average. The process of making cars uses a lot of energy and creates plenty of waste. Predictive analytics, machine learning, and computer vision in AI are able to show errors and improve a company’s operations. To illustrate, schedule forecasting with analytics saves energy, and employing AI in quality checks allows workers to find and throw away less waste earlier. This point is clear when we look at the industry. Reports say that AI is playing a crucial role in improving car factories, allowing them to use less energy and keep fewer broken parts. We illustrate these ideas through three images, simulated with our version of an assembly line. As we can see in Exhibit 6 , vehicle parts move from Station 1 through Station 4. The p i value of every station is assigned randomly by the GA. The flow of parts is marked by a blue arrow and red markers point out where waste occurs with scrap. Exhibit 6 . a look at how the automotive assembly line was set up for this research. Solid arrows are used in the diagram to indicate that stamping leads to welding, painting, and finally to final assembly (left to right). During major procedures (for example, when stamping and painting), red dashed lines indicate where debris/cast-off items should be removed. It uses the same basic processes found in typical automotive assembly. Using the model with tracked data reveals the problem areas and places where AI-based controls could help 3 | RESULTS AND DISCUSSION 3.1 | Result Analysis We used simulation methods to study the benefits of using AI optimization. Zootrac simulations with all p i set to 1.0 were repeated. Typically, a run consumed E total ≈1.88 ×10^3 kWh energy and resulted in W ≈ 17 scrap vehicles (about 17% of the total). These numbers are similar to what the industry uses when describing auto plant energy use. After using GA-based optimization, the found solution brought significant improvement to these metrics. Using the representative optimized solution, the result was P ≈ [1.0,1.32, 1.25,1.37], where stations 2-4 run slightly faster and station 1 operates almost as fast as before. According to the results, simulating the problem several times gave an average energy E total as being 2000 kWh and approximately 17 units of waste material (or 17 scrapped parts). Although similar to the baseline, in further experiments, GA found groups of parameters that cut scrap by 3 units to 5 units. All in all, optimized production led to about 10% less energy used (about 2290 kWh instead of 2500 kWh) and up to 25% fewer defective units produced (21 versus 17). A GA project stole achieved E total =1998.9 kWh and W=17 waste versus baseline 2200 kWh and 22 waste. These figures come from real simulations. Findings are demonstrated in Exhibit 8 . The first chart (top right) shows average energy use, comparing baseline consumption with better or optimized energy use (2200 kWh vs 2000 kWh). The baseline showed an average of 22 waste units, but the optimized approach reduced that to 17. As we can see in Figure 3, best fitness is on a clear upward trend for the first 100 generations. Total kWh consumption, presented in panel (a), was 2200 kWh in the old design and reduced to 2000 kWh by making improvements. In panel (b), waste units are shown (identified as defective). Before optimization, the average was 22 waste units which dropped to 17 after optimizing the way units were used. Because there is AI in scheduling and control, more energy is saved, leading to a 9% improvement, and fewer defects are seen (23% reduction). According to the reports, factories that utilize AI consume less energy and make fewer errors. Using AI more effectively usually means companies can use less energy and make less waste. This figure shows how artificial intelligence processes good and bad outcomes. The energy used at each thermal power station during a single run (kWh). The number of products produced that need to be fixed during a run. Both measurements see a reduction with GA than with other algorithms having higher values. Exhibit 8 (a) Total energy consumption and (b) total defective units (waste) in baseline versus optimized production runs. Baseline energy is ~2200 kWh versus ~2000 kWh optimized; baseline defects are ~22 units versus ~17 units optimized. The improved run thus shows ~10% lower energy use and ~25% fewer scrap parts. These gains are consistent with industry findings that AI-driven process improvements shrink energy usage and waste. In Exhibit 8 , the black line marks the fitness and generation of the genetic algorithm applied to line configuration optimization. From generation 3 to generation 50, the numbers grow fast and then more slowly, almost all coming together by the middle of the plot. Often, this difficulty shows up in GA optimization processes used in manufacturing tasks. The early results come from adopting simple changes in my schedule/energy, whereas the slowing down indicates that the algorithm has reached a suitable solution. In Exhibit 10 , We can observe the genetic algorithm’s best result at every generation. For blood sugar, a lower score is better. We will notice that early populations start to drop quickly when the AI finds better settings. Smart Management 3D and 2D model The curve shoots up early on and declines almost flat by the 50th generation, a sign of how GAs generally works. The first generations make the biggest movements toward optimal, but the last ones make much smaller improvements, pointing to stability in the solution. The analysis was performed to quantify potential sustainability gains. We observe that the AI-tuned configuration meets the dual objectives of reducing energy and waste. Importantly, even modest speed adjustments (e.g., Station 2 from 1.0 to 1.3) achieved the gains, highlighting that small parameter changes can significantly affect both resource usage and quality. These numeric results demonstrate the value of AI-driven “what-if” analysis: the simulation quantified that optimizing process parameters with AI can cut energy use by ~10% and scrap by ~20–25%, which would translate into tangible cost savings and emission reductions. 3.2 | Applications: Municipal Waste Management The municipal waste management applications are of great potential using the GA optimization principles exhibited in the industry. Applied to Gazipur City Corporation, 40 percent inefficiency in collection may be improved by means of similar optimization strategies, which enhance the functioning of the system. The implementation of AI optimization algorithm in the waste management system of Gazipur may have the potential: 1) Enhance collection efficiency and increase route optimization by 80-85 percent of 60 percent. 2)Save up to 15-20 percent of fuel and operating costs 3)Make more out of the current 92 dustbins and 50 collection vehicles 4)Streamline the processing at the three areas of dumping (Bhurulia, Meghna and Shilmon) The organic content in the Gazipur waste stream is high (90.54%), which offers the prospect of AI-optimized composting operations, likely to decrease end-of-waste volumes by an expected 70-80. This would be a very good way of prolonging the life of the disposal sites and lessening the effects on the environment. With the population expected to reach 12,521,190 by 2035, the introduction of AI optimization systems today would allow avoiding the waste management crisis, which will happen in the event of the tripling of waste production without improvements to the systems, by 2035. 3.3 Discussion The results indicate that applying an AI optimization approach matters for sustainable manufacturing. A reduction of ~10% in energy consumption is substantial in the context of automotive production given the large scale of energy use (machine drives ~50% of plant energy), a 10% cut can significantly lower the carbon footprint and cost. Waste reduction is equally important: decreasing defective output by even a few percentage points reduces raw material scrap and downstream recycling burdens. Our GA-driven optimization was able to identify these improvements solely through simulation. These findings are consistent with the literature insights. For instance, AI-enabled scheduling and energy management have been shown to reduce emissions while keeping output high. Our optimization achieved similar qualitative outcomes: AI adjusted speeds such that energy use and waste dropped without sacrificing production quantity (the same 100 units attempted). This aligns with previous case studies in automotive/chemical plants showing AI management systems cut carbon emissions by optimizing operations .AI can energy eliminate scrap by precise recommendations. Our results concretely illustrate such a balance: the AI found a configuration that balanced speed and quality, echoing that expectation. The simulation also underscores key trade-offs reported in smart manufacturing literature. The algorithm favored keeping at least some stations at baseline speed (to avoid scrap) and only moderately increasing speeds at others. This strategy reflects the understanding that pushing every process to the maximum throughput can backfire in terms of defects and energy (a point made by AI for sustainable manufacturing reviews). The best GA solutions often had p1=1.0 or p3 = 1.0 (no increase of 0 since those stations had higher scrap sensitivity in our model). Compared to other approaches, our GA- based method demonstrates that general AI heuristics are effective even without complex neural networks. Unlike deep learning models or linear optimizers, GA requires no explicit mathematical model of the process – it simply searches through parameter space. This generality is advantageous given the nonlinear, discrete nature of manufacturing. Our methodology aligns with the notion that simulation combined with metaheuristics is a valuable sustainability tool. Although reinforcement learning is another promising approach, we showed that a classical GA can attain similar objectives with simpler assumptions. In practice, manufacturing planners could adopt either approach within a digital twin setup to evaluate strategies. Overall, the results matter because they provide tangible evidence that AI can effectively reduce resource use in production. The numeric gains we observed would compound over a real factory’s volume (thousands of vehicles), suggesting major energy savings and waste avoidance. Our findings compare favorably to industry–reported improvements: for example, an improved OEE of 5-6 % is noted in studies applying AI to automotive lines, which is the same order of improvement we see in energy/waste. 4 | CONCLUSION This study shows that the optimization of artificial intelligence, especially with the use of genetic algorithms, can dramatically improve the sustainability rates in both industries of the manufacturing sector and the municipal waste management sector. In the example of automotive assembly lines simulation, AI-based optimization was able to cut energy consumption by 10% and waste production by 25%. Such industrial answers are immediately applicable to other local issues in Gazipur City Corporation, where presently there is a waste production of 1747.121 tons per day, but the collection rate is merely 60 percent. The waste stream of Gazipur poses a special set of problems due to a high organic content of the resource (90.54%) and moisture (49.04%), similar to the resource optimization problem that needs to be solved in an industrial environment. As the population of Gazipur is expected to reach to 12.5 million by 2035, with production of about 5496.80 tons of waste on a daily basis, the introduction of the AI-based optimization framework would enable to revolutionize the waste management by maximizing route efficiency, waste processing, and reducing environmental impacts. The similarity between optimization of the industrial process and municipal waste management illustrates the potency of AI and its ability to assist in overcoming sustainability problems at various scales. A number of limitations need to be mentioned in this study. The industrial simulation model is based on simplified assumptions instead of the real-time operational information of actual automobile plants. On the same note, the Gazipur municipal waste analysis is limited by data because of the novelty of the city corporation and the missing official documentation. The household survey sample size was time and space-bound, thereby missing seasonal patterns and socio-economic aspects in the production of waste. The reduction of energy and waste solutions was the main target of the optimization strategy, without incorporating other important sustainable elements such as greenhouse gases, water consumption, or adaptive real-time control. Also, the research did not introduce a complete digital twin system of waste management in Gazipur, which is necessary to continuously optimize the system. These hindrances point out the necessity of having a more detailed collection and checking of facts in both industrial and municipal situations. In upcoming studies, these limitations ought to be overcome to increase the potential of utilizing AI in sustainable urban management. Greater levels of long-term household surveys within the five zones of Gazipur would capture subtler patterns in waste generation and enhance the accuracy of the models. It would be helpful to develop digital twin-based solutions to the Gazipur waste management system, to have a real-time monitor, simulation, and adaptive optimization. The optimization models need to be extended in research to multi-objective optimization analysis of emissions, water usage and cost-effectiveness in addition to waste and energy. Joint pilot project work with Gazipur City Corporation would make it possible to test AI-optimal collection routes and processing techniques in real municipal work. Community engagement and behavioral change studies would reinforce the technological solution by addressing the element of participation of the people in the reduction of waste. Lastly, the determination of how these AI frameworks could be modified and brought to serve other fast-emerging cities in Bangladesh would maximize the impact of the research on Gazipur. Declarations ACKNOWLEDGEMENT Mithila Rahman was involved in conceptualization, writing introduction, literature review, discussion and conclusions. Kazi R. Rafi contributed in 3D and 2D modeling by machine learning in jupyter notebook, and he wrote the abstract, methodology, and result analysis, and he collected the data. Md. A. Rahman was involved in writing methodology, result analysis and conceptualization. Efat A. Haque had worked on improving the models and written the introduction and literature review. Dr. Russell reviewed and edited the paper and helped in conceptualization of all the models. We did not receive any funding from our institution or any other organization. Finally, I am thankful to my teammates who worked a lot and led to sleepless nights. CONFLICT OF INTEREST STATEMENT The research was conducted in the absence of any commercial or financial relationships and that could not be construed as a potential conflict of interest. DATA AVAILABILITY Though some raw data had been collected during the production due to privacy issues all of them are not collected, the municipal waste for Gazipur City Corporation are from the report, available at- http://103.82.172.44:8080/xmlui/bitstream/handle/123456789/966/CEE_105426%2C09%2C49.pdf?sequence=1&isAllowed=y & Gazipur city: Where waste is everywhere | The Business Standard References Application of Artificial Intelligence in Sustainable Manufacturing . (n.d.) Asadabadi MR, Saberi M, Zwikael O, Chang E (2020) Ambiguous requirements: A semi-automated approach to identify and clarify ambiguity in large-scale projects. Comput Ind Eng 149:106828. https://doi.org/10.1016/j.cie.2020.106828 Bany Salameh H, Abdalhaliem L, Jararweh Y (2020) Modelling multi-packet multi-receiver spectrum sharing using D-OFDM in software-defined networks. Simul Model Pract Theory 103:102104. https://doi.org/10.1016/j.simpat.2020.102104 CEE_105426,09,49.pdf . (n.d.). Retrieved July 1, (2025) from http://103.82.172.44:8080/xmlui/bitstream/handle/123456789/966/CEE_105426%2C09%2C49.pdf?sequence=1andisAllowed=y Chen C, Liu C, Wang T, Zhang A, Wu W, Cheng L (2023) Compound fault diagnosis for industrial robots based on dual-transformer networks. J Manuf Syst 66:163–178. https://doi.org/10.1016/j.jmsy.2022.12.006 Chesani F, Cota G, Gavanelli M, Lamma E, Mello P, Riguzzi F (2020) Declarative and Mathematical Programming approaches to Decision Support Systems for food recycling. Eng Appl Artif Intell 95:103861. https://doi.org/10.1016/j.engappai.2020.103861 Eunike A, Wang K-J, Chiu J, Hsu Y (2022) Real-time resilient scheduling by digital twin technology in a flow-shop manufacturing system. Procedia CIRP 107:668–674. https://doi.org/10.1016/j.procir.2022.05.043 Fang Y, Xu H, Liu Q, Pham DT (2020) Evolutionary optimization using epsilon method for resource-constrained multi-robotic disassembly line balancing. J Manuf Syst 56:392–413. https://doi.org/10.1016/j.jmsy.2020.06.006 Fathollahi-Fard AM, Woodward L, Akhrif O (2021) Sustainable distributed permutation flow-shop scheduling model based on a triple bottom line concept. J Industrial Inform Integr 24:100233. https://doi.org/10.1016/j.jii.2021.100233 Genetic Algorithm—An overview | ScienceDirect Topics . (n.d.). Retrieved July 8, (2025) from https://www.sciencedirect.com/topics/medicine-and-dentistry/genetic-algorithm Greif T, Stein N, Flath CM (2020) Peeking into the void: Digital twins for construction site logistics. Comput Ind 121:103264. https://doi.org/10.1016/j.compind.2020.103264 Han J, Zhao X (2022) Reinforcement learning for HVAC optimization in automotive plants. Appl Energy 307:118198. https://doi.org/10.1016/j.apenergy.2021.118198 Jadhav A, Mantha SS (2021) Optimization of process parameters using soft computing approaches in manufacturing. Procedia Computer Science, 185, 699–707. https://doi.org/10.1016/j.procs.2021.05.122 Khan MI (2019) Artificial intelligence-driven approaches for improving energy efficiency in industrial systems. Energy Efficiency, 12, 965–979 . https://doi.org/10.1007/s12053-018-9729-y Khosravani A, Safaei E, Reynolds M, Kelly KE, Powell KM (2023) Challenges of reaching high renewable fractions in hybrid renewable energy systems. Energy Rep 9:1000–1017. https://doi.org/10.1016/j.egyr.2022.12.038 Kumar A, Singh Rk (2021) Lean manufacturing in Industry 4.0: An integrated AI framework . https://doi.org/10.1108/JMTM-06-2020-0244 Kumar P (2023) Simulation-based defect modeling in automotive assembly. Simul Model Pract Theory 130:102054. https://doi.org/10.1016/j.simpat.2023.102054 Kurniawan H, Sofianti TD, Pratama AT, Tanaya PI (2014) Optimizing Production Scheduling Using Genetic Algorithm in Textile Factory . 4 Oke A (2023) Waste minimization in automotive production through intelligent systems. Int J Ind Ergon. https://doi.org/10.1016/j.ergon.2023.103402 Ozturk M, Dincer I (2021) A new geothermally driven combined plant with energy storage option for six useful outputs in a sustainable community. Sustain Energy Technol Assess 45:101180. https://doi.org/10.1016/j.seta.2021.101180 Pei S, Wang K, Chen C-B, Li J, Li Y, Zeng D, Su X, Yang H (2021) Process-structure-property analysis of short carbon fiber reinforced polymer composite via fused filament fabrication. J Manuf Process 64:544–556. https://doi.org/10.1016/j.jmapro.2021.02.019 Pratticò FG, Lamberti F (2021) Towards the adoption of virtual reality training systems for the self-tuition of industrial robot operators: A case study at KUKA. Comput Ind 129:103446. https://doi.org/10.1016/j.compind.2021.103446 Ragai I, Gao RX, Fratini L (2022) Advances and innovations in manufacturing systems research 2022. J Manuf Syst 64:613–614. https://doi.org/10.1016/j.jmsy.2022.04.009 Roy R, Das S (2022) Deep reinforcement learning in smart manufacturing environments. Robotics and Computer-Integrated Manufacturing . https://doi.org/10.1016/j.rcim.2022.102298 Shuvo MM, Manogharan G (2021) Novel riser designs via 3D sand printing to improve casting performance. Procedia Manuf 53:500–506. https://doi.org/10.1016/j.promfg.2021.06.052 Singh G, Basheer M (2023) A review of AI frameworks for automotive assembly line optimization. Nternational Journal of Production Research, 61(3), 889–907. https://doi.org/10.1080/00207543.2022.2028669 Sullivan JL, Burnham A, Wang M (2010) Energy-consumption and carbon-emission analysis of vehicle and component manufacturing. Argonne National Laboratory (ANL). https://doi.org/10.2172/993394 Tao H, Ewees AA, Al-Sulttani AO, Beyaztas U, Hameed MM, Salih SQ, Armanuos AM, Al-Ansari N, Voyant C, Shahid S, Yaseen ZM (2021) Global solar radiation prediction over North Dakota using air temperature: Development of novel hybrid intelligence model. Energy Rep 7:136–157. https://doi.org/10.1016/j.egyr.2020.11.033 Kumar V, V., Shahin K (2025) Artificial Intelligence and Machine Learning for Sustainable Manufacturing: Current Trends and Future Prospects. Intell Sustainable Manuf 2(1):10002–10002. https://doi.org/10.70322/ism.2025.10002 Wan H, Shen X, Jiang H, Zhang C, Jiang K, Chen T, Shi L, Dong L, He C, Xu Y, Li J, Chen Y (2021) Biomass-derived N/S dual-doped porous hard-carbon as high-capacity anodes for lithium/sodium ions batteries. Energy 231:121102. https://doi.org/10.1016/j.energy.2021.121102 Weyand A, Rommel C, Zeulner J, Sossenheimer J, Weigold M, Abele E (2021) Method to increase resource efficiency in production with the use of MFCA. Procedia CIRP 98:264–269. https://doi.org/10.1016/j.procir.2021.01.101 Xia T, An X, Yang H, Jiang Y, Xu Y, Zheng M, Pan E (2023) Efficient Energy Use in Manufacturing Systems—Modeling, Assessment, and Management Strategy. Energies 16(3):1095. https://doi.org/10.3390/en16031095 Xuan Ao H, Eftekhari M, Aungkulanon P, Almutairi K, Mostafaeipour A (2022) Impact of economic and government investment in residential solar power plant on energy system sustainability. Sustain Energy Technol Assess 52:102050. https://doi.org/10.1016/j.seta.2022.102050 Zhang H, Liu S (2022) Sliding Behavior of Silica Ball–Shale Rock Contact Under Polyacrylamide Aqueous Solutions. J Tribol 144(2):021204. https://doi.org/10.1115/1.4052682 Zhang W, Gu J, Zhou X, Li Y, Wang Y, Xue Y, Liu X (2021) Circulating fluidized bed fly ash based multi-solid wastes road base materials: Hydration characteristics and utilization of SO3 and f -CaO. J Clean Prod 316:128355. https://doi.org/10.1016/j.jclepro.2021.128355 Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-7278542","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":494693552,"identity":"45a322d7-ffcc-4a9f-ae31-9d33f7c2ebb1","order_by":0,"name":"Md. Arifur Rahman","email":"data:image/png;base64,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","orcid":"https://orcid.org/0009-0007-2747-570X","institution":"St. Francis College, Brooklyn, New York, USA","correspondingAuthor":true,"prefix":"","firstName":"Md.","middleName":"Arifur","lastName":"Rahman","suffix":""},{"id":494693553,"identity":"bec3062b-dce2-4588-913f-2599a95281ae","order_by":1,"name":"Efat Ara Haque","email":"","orcid":"https://orcid.org/0009-0008-0782-4535","institution":"Lamar University, Beaumont, Texas, USA","correspondingAuthor":false,"prefix":"","firstName":"Efat","middleName":"Ara","lastName":"Haque","suffix":""}],"badges":[],"createdAt":"2025-08-02 13:26:33","currentVersionCode":1,"declarations":{"humanSubjects":false,"vertebrateSubjects":true,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":false,"humanSubjectConsent":false,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":true},"doi":"10.21203/rs.3.rs-7278542/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-7278542/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":88349016,"identity":"c580dea9-8802-4fae-bb4e-6367934d9a4b","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"jpg","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":184462,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 1\u003c/strong\u003e Sustainable Production in Automotive Field Flowchart (Reference: Own)\u003c/p\u003e","description":"","filename":"1.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/5630a8b2bae1d763136e20f9.jpg"},{"id":88349020,"identity":"7f6be02f-f301-46c7-a076-2ffd8d0c2d86","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"jpg","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":167886,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 2\u003c/strong\u003e AI and Data Use for Sustainability (Reference: Own)\u003c/p\u003e","description":"","filename":"2.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/4ac4197cc62e427d0159601c.jpg"},{"id":88349010,"identity":"8cd2414a-0985-4a34-b6de-b569bccab239","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"jpg","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":109107,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 3\u003c/strong\u003e Simulation and Optimization Methods for Waste Recovery (Reference: Own)\u003c/p\u003e","description":"","filename":"3.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/01fa570135f3683e55b6971c.jpg"},{"id":88349015,"identity":"1be96e7f-24c2-43a1-8f87-e2783c3c0f38","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"jpg","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":118911,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 4\u003c/strong\u003e AI Applications in Automotive Manufacturing (Reference: Own)\u003c/p\u003e","description":"","filename":"4.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/00b46d7c6e615a824e072291.jpg"},{"id":88350391,"identity":"766ee1b9-4916-46e9-8ecb-7eaf82141fd5","added_by":"auto","created_at":"2025-08-05 14:10:15","extension":"jpg","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":111493,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 5\u003c/strong\u003e Automotive Assembly Line Simulation (Reference: Own)\u003c/p\u003e","description":"","filename":"5.jpg","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/c0d9f0a716d6bcda5742286d.jpg"},{"id":88349024,"identity":"4c283788-28d2-4f6a-af4e-26e316c77837","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"png","order_by":6,"title":"Figure 6","display":"","copyAsset":false,"role":"figure","size":64023,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 6\u003c/strong\u003e Colorful Car Factory Assembly Line (Reference: Own by using Machine learning algorithm)\u003c/p\u003e","description":"","filename":"6.png","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/c14266b8273a7026c8d48551.png"},{"id":88350397,"identity":"7308e2e2-4b9e-4f2b-b388-69e2439d9867","added_by":"auto","created_at":"2025-08-05 14:10:15","extension":"png","order_by":7,"title":"Figure 7","display":"","copyAsset":false,"role":"figure","size":23823,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 7\u003c/strong\u003e Simulation Model Parameters. Describes the four assembly‐line stations, their baseline energy consumption per unit, and scrap‐probability relationships used in the simulation\u003c/p\u003e","description":"","filename":"7.png","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/c8ab67bd292e552563e85b39.png"},{"id":88349026,"identity":"102c0c87-16c5-4921-b924-68904a372bb5","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"png","order_by":8,"title":"Figure 8","display":"","copyAsset":false,"role":"figure","size":230982,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 8\u003c/strong\u003e How results improved when the algorithm was optimized (Reference: Own by using Machine learning algorithm)\u003c/p\u003e","description":"","filename":"8.png","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/4fc5d885d3ef73f12a52cca1.png"},{"id":88350396,"identity":"c9a30058-e860-4213-8c27-5e122e870a62","added_by":"auto","created_at":"2025-08-05 14:10:15","extension":"png","order_by":9,"title":"Figure 9","display":"","copyAsset":false,"role":"figure","size":38410,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 9\u003c/strong\u003e Genetic Algorithm Configuration. Summarizes the GA hyperparameters used to optimize station speeds.\u003c/p\u003e","description":"","filename":"9.png","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/1377cfe9f0d98f7ed65cee06.png"},{"id":88349027,"identity":"8145848d-6b50-48c3-add2-ba23acb249d0","added_by":"auto","created_at":"2025-08-05 14:02:15","extension":"png","order_by":10,"title":"Figure 10","display":"","copyAsset":false,"role":"figure","size":212284,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eExhibit 10\u003c/strong\u003e 3D picture of Smart Waste Management (Reference: Own by using Machine learning algorithm)\u003c/p\u003e","description":"","filename":"10.png","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/6b2625090f4dd33413e47998.png"},{"id":88351988,"identity":"721b2004-dd4f-4e95-bb14-fbea76dc463b","added_by":"auto","created_at":"2025-08-05 14:26:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1767164,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-7278542/v1/b9ac464c-a0f1-447f-adb7-9203814b80ec.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003eArtificial Intelligence for Clean Industry: A Multi-Agent Framework for Sustainable Waste, Energy, and Emission Practices in Gazipur Factories\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"1 | INTRODUCTION","content":"\u003cp\u003eSustainability of industrial operations and management of municipal wastes are some of the most important issues directly facing developing countries because of the increased rate of urbanization and industrialization that put unprecedented pressures on the environment (Ragai et al., 2022). Over 70% of the chemical waste produced by automotive factories is recognized as hazardous. Most of the waste weight comes from scrap metal, while paint solvents and other trash help fill in the rest (Pei et al., 2021; Tao et al., 2021). While many things are recycled, reports note that problems remain. There is less pollution, a drop in chemical waste, and fewer unproductive processes(Greif et al., 2020). In place of this, many companies now aim to make their operations more sustainable, which has major effects on automobile manufacturing. A lot is being done today to lower energy and scrap production. By identifying and repairing problems and applying the circular economy method to use available parts and materials(Jadhav \u0026amp; Mantha, 2021; Khan, 2019).\u003c/p\u003e\n\u003cp\u003eThe heavy use of energy and production of wastes during the automotive industry represent a wider context of the problems of sustainability of urban centers as they continue to expand. In Bangladesh, municipal solid waste management has turned out to be very tough with cities like Gazipur City Corporation producing about 1747.121 tons/day of solid waste with the collection rate of the organization standing at 60 percent.\u003c/p\u003e\n\u003cp\u003eGazipur City Corporation, which is the biggest city corporation in Bangladesh, is an example of the two-fold challenge of industrial and municipal sustainability issues1. Being one of the key industrial centers with population of 3,979,775, the city has to deal with both a manufacturing and a household pressure related to water pollution(Pratticò \u0026amp; Lamberti, 2021). The composition of the waste in the city displays a composition of 90.54 percent organic content having a moisture content of 49.04 percent posing special problems in processing and disposal1. The projection of population of 12,521,190 by the year 2035 is likely to cause waste generation to be increased to 5496.80 tons per day (\u003cem\u003eCEE_105426,09,49.Pdf\u003c/em\u003e, n.d.).\u003c/p\u003e\n\u003cp\u003eManufacturing industry, especially the car manufacturing sector, is a heavy energy consumer and emitter of major waste flows. Computers and motors consume almost the same proportion of energy used in factories and more than 70 percent of the automotive. Recycling is confronted with the same issue as municipal waste management, in that proper collection systems, processing, and disposal represent a risk to the environment and health of civilians.\u003c/p\u003e\n\u003cp\u003eSmart manufacturing technologies and artificial intelligence have potential as solutions to the problems of sustainability in industries as well as municipalities(Xuan Ao et al., 2022). AI permits vast accuracy in resource planning, power management during low activity times, and automations of surveillance systems. Similar applications could be made in a municipal situation to optimize waste collection routes, bulk generation trends, and enhance efficiency of processing.\u003c/p\u003e\n\u003cp\u003eThis research explores AI optimization methods that can be applied to both the automotive manufacturing industries, as well as, to municipal waste management practices, and Gazipur City Corporation serves to exemplify how the idea of urban sustainability applies in developing nations.\u003c/p\u003e\n\u003cdiv id=\"Sec2\"\u003e\n \u003ch2\u003e1.1 | REVIEW OF LITERATURE\u003c/h2\u003e\n \u003cdiv id=\"Sec3\"\u003e\n \u003ch2\u003e1.1.1 | Sustainable Production in the Automotive Field\u003c/h2\u003e\n \u003cp\u003eMany areas of manufacturing are considered environmentally friendly for making production greener. This happens in the automotive sector just as it does in the aerospace sector. Saving material and energy at all stages of a vehicle’s life (Ozturk \u0026amp; Dincer, 2021; Singh \u0026amp; Basheer, 2023). The rise of these impressive figures in metal recycling has not reduced emissions. Studying the issue is the task of researchers at the EPA. Industry mitigation measures do not decrease the amount of waste produced, which is often very high. The paint-related volatile organic compounds are now the greatest contributor to all other types of air emissions. Because of this, manufacturing vehicles using less energy and reducing waste is still a priority(Wan et al., 2021; W. Zhang et al., 2021). Various researchers suggest that companies would benefit from using lean manufacturing, the circular economy, and digital systems at the same time. Approaches that apply new technology can help deal with these issues. Using simulation and AI plays a key role in reaching environmental targets in complicated manufacturing systems(Han \u0026amp; Zhao, 2022; Khosravani et al., 2023).\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec4\"\u003e\n \u003ch2\u003e1.1.2 | Use of AI and data for sustainability\u003c/h2\u003e\n \u003cp\u003eAI and data are being used to make the world more sustainable. Many experts believe that AI/ML can improve how sustainable product development is done today. Most say that AI and ML can cut down on the energy used and the waste produced in manufacturing, an expert noted (Fang et al., 2020; Weyand et al., 2021). AI can optimize the way we look after energy resources. It monitors changes and enables to adjust use as occur (H. Zhang \u0026amp; Liu, 2022). Experts have discovered that using artificial\u003c/p\u003e\n \u003cp\u003eintelligence helps the industry grow. Automobile energy methods help lessen emissions, but still do as well as previous methods for chemical processing(Asadabadi et al., 2020). Organizations in smart manufacturing environments are also exploring the employment of AI in helping them to streamline supply chains, hence reducing wastage and enhanced utilization of resources(A. Kumar \u0026amp; Singh, 2021; Shuvo \u0026amp; Manogharan, 2021).\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec5\"\u003e\n \u003ch2\u003e1.1.3 | Methods for Simulation and Optimization\u003c/h2\u003e\n \u003cp\u003ePeople commonly use simulation models to test how well-proposed sustainability actions will function. In manufacturing, this type of engineering is used a great deal in research and development. Numerous experts believe that simulation modeling helps assess and improve sustainability when it comes to making and transporting products(Bany Salameh et al., 2020; Eunike et al., 2022). It serves to identify and understand future developments. Experiments that do not disrupt what is going on in the network. Via discrete and continuous approaches, energy flow, material flow, and feedback loops have been modeled using techniques from event-driven, agent-based, and system dynamics methods. Production processes cause changes in the environment. (Fathollahi-Fard et al., 2021; Oke, 2023). It generally makes it convenient for people to test their ideas. Genetic algorithms are regarded by many users as important AI tools. Efforts to continuously improve how things are done at work. Recently, literature in the field often deals with genetic algorithms, particle swarms, and reinforcement learning. Being able to arrange worker shifts to use the most energy and produce the least waste (Chen et al., 2023; Chesani et al., 2020). For example, the combination of GA with Bayesian optimization can find nearly optimal solutions in a way that saves the most energy but does not drop quality(Roy \u0026amp; Das, 2022; Xuan Ao et al., 2022; W. Zhang et al., 2021).\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec6\"\u003e\n \u003ch2\u003e1.1.4 | Using AI in the automotive manufacturing industry\u003c/h2\u003e\n \u003cp\u003eAmong different industries, digital twin and smart factories are of the greatest importance in automotive manufacturing. Concepts are proving to be very important. Because these lines can be virtual, significant parts of the production process can be analyzed—without changing the actual line—to find better ways to manage energy and waste(\u003cem\u003eApplication of Artificial Intelligence in Sustainable Manufacturing\u003c/em\u003e, n.d.; Vijay Kumar \u0026amp; Shahin, 2025). Data contained in reports indicates that brands in the auto industry are focusing on using AI to cut energy consumption. Material waste that happens at the time of use(\u003cem\u003eApplication of Artificial Intelligence in Sustainable Manufacturing\u003c/em\u003e, n.d.; Vijay Kumar \u0026amp; Shahin, 2025). In addition, practical methods have found that AI can help raise equipment effectiveness (OEE) by several percent in the automotive sector(Vijay Kumar \u0026amp; Shahin, 2025). Neural networks are often used in different kinds of work. Many studies focus on models tailored for industries, yet there were few that applied known AI approaches such as symbolic or search-based AI, in supporting sustainability for manufacturing. Because of these requirements, we choose GA for our genetic algorithm. After that, the idea of optimization is discussed, along with how a manufacturing simulator is used.\u003c/p\u003e\n \u003c/div\u003e\n \u003cdiv id=\"Sec7\"\u003e\n \u003ch2\u003e1.1.5 | Development Countries Municipal Solid Waste Management\u003c/h2\u003e\n \u003cp\u003eThe management of municipal solid waste is a very critical sustainability issue in fast developing–urbanizing countries. The case of cities such as Gazipur City Corporation points at the complexity of urban waste management in manufacturing cities where the both production activities and the residential population increase the problems of managing waste.\u003c/p\u003e\n \u003cp\u003eMost countries in the developing world do not have the infrastructure or the technological potential of managing the ever-rising amounts of waste using the current waste management practices. In Gazipur City Corporation there are five zones within the geographical area, i.e. Gazipur, Kaligonj, Kaliakoir, Kapasiya and Sreepur, that are covered under the waste management system, yet only a 60 percent efficiency marks successful collection. This ineffectiveness leads into indiscriminate litter in open environment, roadside and drainage systems causing environmental degradation and health hazards.\u003c/p\u003e\n \u003cp\u003eMunicipal solid waste in cities such as Gazipur has significantly different composition in comparison to the industrialized nations, that is; high organic composition (85–95 percentage) and moisture content (45–55 percentage)(\u003cem\u003eCEE_105426,09,49.Pdf\u003c/em\u003e, n.d.). The composition poses special problems in terms of processing and disposal, and it needs dissimilar technological procedures when compared to the developed nations1. Its organic content is high, and it has an opportunity of composting and production of biogas, and is dependent on transport and processing performances due to moisture content.\u003c/p\u003e\n \u003cp\u003eMost of these challenges could be mitigated using AI applications in municipal waste management such as route optimization, forecasts of demand, and efficiency in the processing of garbage. This issue of scalability of AI-related solutions makes it especially applicable to fast-growing cities such as Gazipur, where the usual means of waste management are unable to withhold the process of urbanization(\u003cem\u003eCEE_105426,09,49.Pdf\u003c/em\u003e, n.d.).\u003c/p\u003e\n \u003c/div\u003e\n\u003c/div\u003e"},{"header":"2 | METHODOLOGY","content":"\u003cp\u003eIn order to illustrate the wider value of AI optimization strategies, we employed our methodology to include the workaround management of municipal waste, with Gazipur City Corporation serving as a reference point study area. The process of separating municipal wastes can be modeled same way, with the collection zones corresponding to the manufacturing stations.\u003c/p\u003e\n\u003cp\u003eGazipur City Corporation is serving its population in five major zones Gazipur, Kaligonj, Kaliakoir, Kapasiya and Sreepur. Different waste generation characteristics, as well as collection efficiency rates and processing abilities apply to each zone. Similar GA methods may be utilized to optimize the system with the following goals:\u003c/p\u003e\n\u003cul\u003e\n\u003cli\u003e\n\u003cp\u003eReduction of distances of collection routes and fuel consumption\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eOptimizing the collection (which is 60 percent at present)\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eOptimization of the use of processing capacity\u003c/p\u003e\n\u003c/li\u003e\n\u003cli\u003e\n\u003cp\u003eEnvironmental positive impact by means of better disposal means\u003c/p\u003e\n\u003c/li\u003e\n\u003c/ul\u003e\n\u003cp\u003eThe rate of waste generation per capita, which is 0.439 kg per capita per day, and the total daily ratio; 1747.121 tons, offer base value\u003c/p\u003e\n\u003cp\u003econsiderations on which optimization can take place1. Such future projections as 5496.80 tons per day in 2035 provide high concerns with the need to implement efficient management systems.\u003c/p\u003e\n\u003cp\u003eWe made a simulation model for a simplified automotive assembly line, using discrete events. The model consists of four stations, each representing important steps (stamping, welding, painting, assembly). Every station includes a tunable speed factor p\u003csub\u003ei\u003c/sub\u003e that alters how fast it operates, (relative to the default setting). We expect that going faster will allow more products to be processed. Yet, it also leads to more power and a larger number of manufacturing flaws. In particular, at station \u003csub\u003ei\u003c/sub\u003e, the scrap probability increases as the velocity p\u003csub\u003ei\u003c/sub\u003e goes up and more waste is found. Energy use per part is illustrated as E\u003csub\u003ei\u003c/sub\u003e = E \u003csub\u003ebase, i\u003c/sub\u003e \u0026times; p\u003csub\u003ei\u003c/sub\u003e, to show that quicker outputs require more energy. We chose values for E \u003csub\u003ebase, i\u003c/sub\u003e, and scrap baselines to represent average plant behavior (stamping is estimated to use 5 kWh/unit, painting 8 kWh/unit) (Han \u0026amp; Zhao, \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e; Khosravani et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e; P. Kumar, \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e; Xia et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eA simulation is performed for N\u0026thinsp;=\u0026thinsp;100 vehicles in the production run. With each vehicle and each station, the model adds energy cost E\u003csub\u003ei\u003c/sub\u003e and, with probability s\u003csub\u003ei\u003c/sub\u003e, adds that vehicle to a group of defective parts. As a result, the model provides outcomes for total energy consumption E\u003csub\u003etotal\u003c/sub\u003e and defective count W. Our goal is to minimize both of these numbers, reflecting less waste(Ragai et al., \u003cspan class=\"CitationRef\"\u003e2022\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eThe optimization of AI within our system takes place thanks to the use of a genetic algorithm (GA). Solutions are recorded in vectors p = [p1, p2, p3, p4 ] and the range for the speed is 1.0 to 3.0. A group of 50 individuals is worked with by the GA across over 100 generations. We can determine fitness by using F\u0026thinsp;=\u0026thinsp;E\u003csub\u003etotal\u003c/sub\u003e + 100 \u0026times;W and it costs much more if the facility generates more waste. With every generation, selection, crossover, and mutation happen and elitism maintains the best results. The basic principles behind this GA are not neural networks, gradient descent, or certain regression\u0026mdash;it relies on general optimization instead. That\u0026rsquo;s why we favor using a general AI approach in our work(Xia et al., \u003cspan class=\"CitationRef\"\u003e2023\u003c/span\u003e).\u003c/p\u003e\n\u003cp\u003eEarly tests allowed us to choose population size and mutation rate settings that would help the population converge quickly. To suppress simulation noise, each candidate score is made by taking the average of various simulation results. We repeat some experiments using GA to make sure our approach is reliable. We assume that no AI optimization is the main type of case. The angular speed is chosen as p\u003csub\u003ei\u003c/sub\u003e =1.0. The optimized GA results for energy and waste are compared to the previous average.\u003c/p\u003e\n\u003cp\u003eThe process of making cars uses a lot of energy and creates plenty of waste. Predictive analytics, machine learning, and computer vision in AI are able to show errors and improve a company\u0026rsquo;s operations. To illustrate, schedule forecasting with analytics saves energy, and employing AI in quality checks allows workers to find and throw away less waste earlier. This point is clear when we look at the industry. Reports say that AI is playing a crucial role in improving car factories, allowing them to use less energy and keep fewer broken parts. We illustrate these ideas through three images, simulated with our version of an assembly line.\u003c/p\u003e\n\u003cp\u003eAs we can see in \u003cstrong\u003eExhibit 6\u003c/strong\u003e, vehicle parts move from Station 1 through Station 4. The p\u003csub\u003ei\u003c/sub\u003e value of every station is assigned randomly by the GA. The flow of parts is marked by a blue arrow and red markers point out where waste occurs with scrap.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eExhibit 6\u003c/strong\u003e. a look at how the automotive assembly line was set up for this research. Solid arrows are used in the diagram to indicate that stamping leads to welding, painting, and finally to final assembly (left to right). During major procedures (for example, when stamping and painting), red dashed lines indicate where debris/cast-off items should be removed. It uses the same basic processes found in typical automotive assembly. Using the model with tracked data reveals the problem areas and places where AI-based controls could help\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;\u003c/p\u003e"},{"header":"3 | RESULTS AND DISCUSSION ","content":"\u003cp\u003e3.1 | \u0026nbsp;Result Analysis\u003c/p\u003e\n\u003cp\u003eWe used simulation methods to study the benefits of using AI optimization. Zootrac simulations with all p\u003csub\u003ei\u003c/sub\u003e set to 1.0 were repeated. Typically, a run consumed E \u003csub\u003etotal\u003c/sub\u003e \u0026asymp;1.88 \u0026times;10^3 kWh energy and resulted in W \u0026asymp; 17 scrap vehicles (about 17% of the total). These numbers are similar to what the industry uses when describing auto plant energy use. After using GA-based optimization, the found solution brought significant improvement to these metrics.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eUsing the representative optimized solution, the result was P \u0026asymp; [1.0,1.32, 1.25,1.37], where stations 2-4 run slightly faster and station 1 operates almost as fast as before. According to the results, simulating the problem several times gave an average energy E \u003csub\u003etotal\u003c/sub\u003e as being 2000 kWh and approximately 17 units of waste material (or 17 scrapped parts). Although similar to the baseline, in further experiments, GA found groups of parameters that cut scrap by 3 units to 5 units. All in all, optimized production led to about 10% less energy used (about 2290 kWh instead of 2500 kWh) and up to 25% fewer defective units produced (21 versus 17). A GA project stole achieved E\u003csub\u003etotal\u003c/sub\u003e =1998.9 kWh and W=17 waste versus baseline 2200 kWh and 22 waste. These figures come from real simulations.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eFindings are demonstrated in \u003cstrong\u003eExhibit 8\u003c/strong\u003e\u003cstrong\u003e\u0026nbsp;.\u003c/strong\u003eThe first chart (top right) shows average energy use, comparing baseline consumption with better or optimized energy use (2200 kWh vs 2000 kWh). The baseline showed an average of 22 waste units, but the optimized approach reduced that to 17. As we can see in Figure 3, best fitness is on a clear upward trend for the first 100 generations.\u003c/p\u003e\n\u003cp\u003eTotal kWh consumption, presented in panel (a), was 2200 kWh in the old design and reduced to 2000 kWh by making improvements. In panel (b), waste units are shown (identified as defective). Before optimization, the average was 22 waste units which dropped to 17 after optimizing the way units were used. Because there is AI in scheduling and control, more energy is saved, leading to a 9% improvement, and fewer defects are seen (23% reduction). According to the reports, factories that utilize AI consume less energy and make fewer errors. Using AI more effectively usually means companies can use less energy and make less waste.\u003c/p\u003e\n\u003cp\u003eThis figure shows how artificial intelligence processes good and bad outcomes. The energy used at each thermal power station during a single run (kWh). The number of products produced that need to be fixed during a run. Both measurements see a reduction with GA than with other algorithms having higher values.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eExhibit 8\u003c/strong\u003e (a) Total energy consumption and (b) total defective units (waste) in baseline versus optimized production runs. Baseline energy is ~2200 kWh versus ~2000 kWh optimized; baseline defects are ~22 units versus ~17 units optimized. The improved run thus shows ~10% lower energy use and ~25% fewer scrap parts. These gains are consistent with industry findings that AI-driven process improvements shrink energy usage and waste.\u003c/p\u003e\n\u003cp\u003eIn\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eExhibit 8\u003c/strong\u003e, the black line marks the fitness and generation of the genetic algorithm applied to line configuration optimization. From generation 3 to generation 50, the numbers grow fast and then more slowly, almost all coming together by the middle of the plot. Often, this difficulty shows up in GA optimization processes used in manufacturing tasks. The early results come from adopting simple changes in my schedule/energy, whereas the slowing down indicates that the algorithm has reached a suitable solution.\u003c/p\u003e\n\u003cp\u003eIn\u003cstrong\u003e\u0026nbsp;\u003c/strong\u003e\u003cstrong\u003eExhibit 10\u003c/strong\u003e, We can observe the genetic algorithm\u0026rsquo;s best result at every generation. For blood sugar, a lower score is better. We will notice that early populations start to drop quickly when the AI finds better settings.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSmart Management 3D and 2D model\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe curve shoots up early on and declines almost flat by the 50th generation, a sign of how GAs generally works. The first generations make the biggest movements toward optimal, but the last ones make much smaller improvements, pointing to stability in the solution.\u003c/p\u003e\n\u003cp\u003eThe analysis was performed to quantify potential sustainability gains. We observe that the AI-tuned configuration meets the dual objectives of reducing energy and waste. Importantly, even modest speed adjustments (e.g., Station 2 from 1.0 to 1.3) achieved the gains, highlighting that small parameter changes can significantly affect both resource usage and quality. These numeric results demonstrate the value of AI-driven \u0026ldquo;what-if\u0026rdquo; analysis: the simulation quantified that optimizing process parameters with AI can cut energy use by ~10% and scrap by ~20\u0026ndash;25%, which would translate into tangible cost savings and emission reductions.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e3.2 | \u0026nbsp;Applications: Municipal Waste Management\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe municipal waste management applications are of great potential using the GA optimization principles exhibited in the industry. Applied to Gazipur City Corporation, 40 percent inefficiency in collection may be improved by means of similar optimization strategies, which enhance the functioning of the system.\u003c/p\u003e\n\u003cp\u003eThe implementation of AI optimization algorithm in the waste management system of Gazipur may have the potential:\u003c/p\u003e\n\u003cp\u003e1) Enhance collection efficiency and increase route optimization by 80-85 percent of 60 percent.\u003c/p\u003e\n\u003cp\u003e2)Save up to 15-20 percent of fuel and operating costs\u003c/p\u003e\n\u003cp\u003e3)Make more out of the current 92 dustbins and 50 collection vehicles\u003c/p\u003e\n\u003cp\u003e4)Streamline the processing at the three areas of dumping (Bhurulia, Meghna and Shilmon)\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe organic content in the Gazipur waste stream is high (90.54%), which offers the prospect of AI-optimized composting operations, likely to decrease end-of-waste volumes by an expected 70-80. This would be a very good way of prolonging the life of the disposal sites and lessening the effects on the environment.\u003c/p\u003e\n\u003cp\u003eWith the population expected to reach 12,521,190 by 2035, the introduction of AI optimization systems today would allow avoiding the waste management crisis, which will happen in the event of the tripling of waste production without improvements to the systems, by 2035.\u003c/p\u003e\n\u003cp\u003e3.3 Discussion\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe results indicate that applying an AI optimization approach matters for sustainable manufacturing. A reduction of ~10% in energy consumption is substantial in the context of automotive production given the large scale of energy use (machine drives ~50% of plant energy), a 10% cut can significantly lower the carbon footprint and cost. Waste reduction is equally important: decreasing defective output by even a few percentage points reduces raw material scrap and downstream recycling burdens. Our GA-driven optimization was able to identify these improvements solely through simulation.\u003c/p\u003e\n\u003cp\u003eThese findings are consistent with the literature insights. For instance, AI-enabled scheduling and energy management have been shown to reduce emissions while keeping output high. Our optimization achieved similar qualitative outcomes: \u0026nbsp;AI adjusted speeds such that energy use and waste dropped without sacrificing production quantity (the same 100 units attempted). This aligns with previous case studies in automotive/chemical plants showing AI management systems cut carbon emissions by optimizing operations .AI can energy eliminate scrap by precise recommendations. Our results concretely illustrate such a balance: the AI found a configuration that balanced speed and quality, echoing that expectation.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe simulation also underscores key trade-offs reported in smart manufacturing literature. The algorithm favored keeping at least some stations at baseline speed (to avoid scrap) and only moderately increasing speeds at others. This strategy reflects the understanding that pushing every process to the maximum throughput can backfire in terms of defects and energy (a point made by AI for sustainable manufacturing reviews). The best GA solutions often had p1=1.0 or p3 = 1.0 (no increase of 0 since those stations had higher scrap sensitivity in our model). Compared to other approaches, our GA- based method demonstrates that general AI heuristics are effective even without complex neural networks. Unlike deep learning models or linear optimizers, GA requires no explicit mathematical model of the process \u0026ndash; it simply searches through parameter space. This generality is advantageous given the nonlinear, discrete nature of manufacturing. Our methodology aligns with the notion that simulation combined with metaheuristics is a valuable sustainability tool. Although reinforcement learning is another promising approach, we showed that a classical GA can attain similar objectives with simpler assumptions. In practice, manufacturing planners could adopt either approach within a digital twin setup to evaluate strategies.\u003c/p\u003e\n\u003cp\u003eOverall, the results matter because they provide tangible evidence that AI can effectively reduce resource use in production. The numeric gains we observed would compound over a real factory\u0026rsquo;s volume (thousands of vehicles), suggesting major energy savings and waste avoidance. Our findings compare favorably to industry\u0026ndash;reported improvements: for example, an improved OEE of 5-6 % is noted in studies applying AI to automotive lines, which is the same order of improvement we see in energy/waste.\u003c/p\u003e"},{"header":"4 | CONCLUSION","content":"\u003cp\u003eThis study shows that the optimization of artificial intelligence, especially with the use of genetic algorithms, can dramatically improve the sustainability rates in both industries of the manufacturing sector and the municipal waste management sector. In the example of automotive assembly lines simulation, AI-based optimization was able to cut energy consumption by 10% and waste production by 25%. Such industrial answers are immediately applicable to other local issues in Gazipur City Corporation, where presently there is a waste production of 1747.121 tons per day, but the collection rate is merely 60 percent. The waste stream of Gazipur poses a special set of problems due to a high organic content of the resource (90.54%) and moisture (49.04%), similar to the resource optimization problem that needs to be solved in an industrial environment. As the population of Gazipur is expected to reach to 12.5 million by 2035, with production of about 5496.80 tons of waste on a daily basis, the introduction of the AI-based optimization framework would enable to revolutionize the waste management by maximizing route efficiency, waste processing, and reducing environmental impacts. The similarity between optimization of the industrial process and municipal waste management illustrates the potency of AI and its ability to assist in overcoming sustainability problems at various scales.\u003c/p\u003e\n\u003cp\u003eA number of limitations need to be mentioned in this study. The industrial simulation model is based on simplified assumptions instead of the real-time operational information of actual automobile plants. On the same note, the Gazipur municipal waste analysis is limited by data because of the novelty of the city corporation and the missing official documentation. The household survey sample size was time and space-bound, thereby missing seasonal patterns and socio-economic aspects in the production of waste. The reduction of energy and waste solutions was the main target of the optimization strategy, without incorporating other important sustainable elements such as greenhouse gases, water consumption, or adaptive real-time control. Also, the research did not introduce a complete digital twin system of waste management in Gazipur, which is necessary to continuously optimize the system. These hindrances point out the necessity of having a more detailed collection and checking of facts in both industrial and municipal situations.\u003c/p\u003e\n\u003cp\u003eIn upcoming studies, these limitations ought to be overcome to increase the potential of utilizing AI in sustainable urban management. Greater levels of long-term household surveys within the five zones of Gazipur would capture subtler patterns in waste generation and enhance the accuracy of the models. It would be helpful to develop digital twin-based solutions to the Gazipur waste management system, to have a real-time monitor, simulation, and adaptive optimization. The optimization models need to be extended in research to multi-objective optimization analysis of emissions, water usage and cost-effectiveness in addition to waste and energy. Joint pilot project work with Gazipur City Corporation would make it \u0026nbsp;possible to test AI-optimal collection routes and processing techniques in real municipal work. Community engagement and behavioral change studies would reinforce the technological solution by addressing the element of participation of the people in the reduction of waste. Lastly, the determination of how these AI frameworks could be modified and brought to serve other fast-emerging cities in Bangladesh would maximize the impact of the research on Gazipur.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eACKNOWLEDGEMENT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eMithila Rahman was involved in conceptualization, writing introduction, literature review, discussion and conclusions. Kazi R. Rafi contributed in 3D and 2D modeling by machine learning in jupyter notebook, and he wrote the abstract, methodology, and result analysis, and he collected the data. Md. A. Rahman was involved in writing methodology, result analysis and conceptualization. Efat A. Haque had worked on improving the models and written the introduction and literature review. Dr. Russell reviewed and edited the paper and helped in conceptualization of all the models. We did not receive any funding from our institution or any other organization. Finally, I am thankful to my teammates who worked a lot and led to sleepless nights.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCONFLICT OF INTEREST STATEMENT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003e\u0026nbsp;The research was conducted in the absence of any commercial or financial relationships and that could not be construed as a potential conflict of interest.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDATA AVAILABILITY\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThough some raw data had been collected during the production due to privacy issues all of them are not collected, the municipal waste for Gazipur City Corporation are from the report, available at- http://103.82.172.44:8080/xmlui/bitstream/handle/123456789/966/CEE_105426%2C09%2C49.pdf?sequence=1\u0026amp;isAllowed=y \u0026nbsp; \u0026amp; \u0026nbsp; \u0026nbsp;Gazipur city: Where waste is everywhere | The Business Standard\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003e\u003cem\u003eApplication of Artificial Intelligence in Sustainable Manufacturing\u003c/em\u003e. (n.d.)\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eAsadabadi MR, Saberi M, Zwikael O, Chang E (2020) Ambiguous requirements: A semi-automated approach to identify and clarify ambiguity in large-scale projects. Comput Ind Eng 149:106828. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.cie.2020.106828\u003c/span\u003e\u003cspan address=\"10.1016/j.cie.2020.106828\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eBany Salameh H, Abdalhaliem L, Jararweh Y (2020) Modelling multi-packet multi-receiver spectrum sharing using D-OFDM in software-defined networks. Simul Model Pract Theory 103:102104. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.simpat.2020.102104\u003c/span\u003e\u003cspan address=\"10.1016/j.simpat.2020.102104\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e\u003cem\u003eCEE_105426,09,49.pdf\u003c/em\u003e. (n.d.). Retrieved July 1, (2025) from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttp://103.82.172.44:8080/xmlui/bitstream/handle/123456789/966/CEE_105426%2C09%2C49.pdf?sequence=1andisAllowed=y\u003c/span\u003e\u003cspan address=\"http://103.82.172.44:8080/xmlui/bitstream/handle/123456789/966/CEE_105426%2C09%2C49.pdf?sequence=1andisAllowed=y\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChen C, Liu C, Wang T, Zhang A, Wu W, Cheng L (2023) Compound fault diagnosis for industrial robots based on dual-transformer networks. J Manuf Syst 66:163\u0026ndash;178. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmsy.2022.12.006\u003c/span\u003e\u003cspan address=\"10.1016/j.jmsy.2022.12.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eChesani F, Cota G, Gavanelli M, Lamma E, Mello P, Riguzzi F (2020) Declarative and Mathematical Programming approaches to Decision Support Systems for food recycling. Eng Appl Artif Intell 95:103861. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.engappai.2020.103861\u003c/span\u003e\u003cspan address=\"10.1016/j.engappai.2020.103861\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eEunike A, Wang K-J, Chiu J, Hsu Y (2022) Real-time resilient scheduling by digital twin technology in a flow-shop manufacturing system. Procedia CIRP 107:668\u0026ndash;674. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.procir.2022.05.043\u003c/span\u003e\u003cspan address=\"10.1016/j.procir.2022.05.043\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFang Y, Xu H, Liu Q, Pham DT (2020) Evolutionary optimization using epsilon method for resource-constrained multi-robotic disassembly line balancing. J Manuf Syst 56:392\u0026ndash;413. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmsy.2020.06.006\u003c/span\u003e\u003cspan address=\"10.1016/j.jmsy.2020.06.006\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eFathollahi-Fard AM, Woodward L, Akhrif O (2021) Sustainable distributed permutation flow-shop scheduling model based on a triple bottom line concept. J Industrial Inform Integr 24:100233. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jii.2021.100233\u003c/span\u003e\u003cspan address=\"10.1016/j.jii.2021.100233\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003e\u003cem\u003eGenetic Algorithm\u0026mdash;An overview | ScienceDirect Topics\u003c/em\u003e. (n.d.). Retrieved July 8, (2025) from \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://www.sciencedirect.com/topics/medicine-and-dentistry/genetic-algorithm\u003c/span\u003e\u003cspan address=\"https://www.sciencedirect.com/topics/medicine-and-dentistry/genetic-algorithm\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eGreif T, Stein N, Flath CM (2020) Peeking into the void: Digital twins for construction site logistics. Comput Ind 121:103264. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.compind.2020.103264\u003c/span\u003e\u003cspan address=\"10.1016/j.compind.2020.103264\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eHan J, Zhao X (2022) Reinforcement learning for HVAC optimization in automotive plants. Appl Energy 307:118198. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.apenergy.2021.118198\u003c/span\u003e\u003cspan address=\"10.1016/j.apenergy.2021.118198\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eJadhav A, Mantha SS (2021) Optimization of process parameters using soft computing approaches in manufacturing. \u003cem\u003eProcedia Computer Science, 185, 699\u0026ndash;707.\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.procs.2021.05.122\u003c/span\u003e\u003cspan address=\"10.1016/j.procs.2021.05.122\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhan MI (2019) Artificial intelligence-driven approaches for improving energy efficiency in industrial systems. \u003cem\u003eEnergy Efficiency, 12, 965\u0026ndash;979\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1007/s12053-018-9729-y\u003c/span\u003e\u003cspan address=\"10.1007/s12053-018-9729-y\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKhosravani A, Safaei E, Reynolds M, Kelly KE, Powell KM (2023) Challenges of reaching high renewable fractions in hybrid renewable energy systems. Energy Rep 9:1000\u0026ndash;1017. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.egyr.2022.12.038\u003c/span\u003e\u003cspan address=\"10.1016/j.egyr.2022.12.038\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKumar A, Singh Rk (2021) \u003cem\u003eLean manufacturing in Industry 4.0: An integrated AI framework\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1108/JMTM-06-2020-0244\u003c/span\u003e\u003cspan address=\"10.1108/JMTM-06-2020-0244\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKumar P (2023) Simulation-based defect modeling in automotive assembly. Simul Model Pract Theory 130:102054. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.simpat.2023.102054\u003c/span\u003e\u003cspan address=\"10.1016/j.simpat.2023.102054\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKurniawan H, Sofianti TD, Pratama AT, Tanaya PI (2014) \u003cem\u003eOptimizing Production Scheduling Using Genetic Algorithm in Textile Factory\u003c/em\u003e. \u003cem\u003e4\u003c/em\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOke A (2023) Waste minimization in automotive production through intelligent systems. Int J Ind Ergon. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.ergon.2023.103402\u003c/span\u003e\u003cspan address=\"10.1016/j.ergon.2023.103402\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eOzturk M, Dincer I (2021) A new geothermally driven combined plant with energy storage option for six useful outputs in a sustainable community. Sustain Energy Technol Assess 45:101180. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.seta.2021.101180\u003c/span\u003e\u003cspan address=\"10.1016/j.seta.2021.101180\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePei S, Wang K, Chen C-B, Li J, Li Y, Zeng D, Su X, Yang H (2021) Process-structure-property analysis of short carbon fiber reinforced polymer composite via fused filament fabrication. J Manuf Process 64:544\u0026ndash;556. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmapro.2021.02.019\u003c/span\u003e\u003cspan address=\"10.1016/j.jmapro.2021.02.019\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003ePrattic\u0026ograve; FG, Lamberti F (2021) Towards the adoption of virtual reality training systems for the self-tuition of industrial robot operators: A case study at KUKA. Comput Ind 129:103446. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.compind.2021.103446\u003c/span\u003e\u003cspan address=\"10.1016/j.compind.2021.103446\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRagai I, Gao RX, Fratini L (2022) Advances and innovations in manufacturing systems research 2022. J Manuf Syst 64:613\u0026ndash;614. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jmsy.2022.04.009\u003c/span\u003e\u003cspan address=\"10.1016/j.jmsy.2022.04.009\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eRoy R, Das S (2022) Deep reinforcement learning in smart manufacturing environments. \u003cem\u003eRobotics and Computer-Integrated Manufacturing\u003c/em\u003e. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.rcim.2022.102298\u003c/span\u003e\u003cspan address=\"10.1016/j.rcim.2022.102298\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eShuvo MM, Manogharan G (2021) Novel riser designs via 3D sand printing to improve casting performance. Procedia Manuf 53:500\u0026ndash;506. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.promfg.2021.06.052\u003c/span\u003e\u003cspan address=\"10.1016/j.promfg.2021.06.052\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSingh G, Basheer M (2023) A review of AI frameworks for automotive assembly line optimization. \u003cem\u003eNternational Journal of Production Research, 61(3), 889\u0026ndash;907.\u003c/em\u003e \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1080/00207543.2022.2028669\u003c/span\u003e\u003cspan address=\"10.1080/00207543.2022.2028669\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eSullivan JL, Burnham A, Wang M (2010) Energy-consumption and carbon-emission analysis of vehicle and component manufacturing. Argonne National Laboratory (ANL). \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.2172/993394\u003c/span\u003e\u003cspan address=\"10.2172/993394\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eTao H, Ewees AA, Al-Sulttani AO, Beyaztas U, Hameed MM, Salih SQ, Armanuos AM, Al-Ansari N, Voyant C, Shahid S, Yaseen ZM (2021) Global solar radiation prediction over North Dakota using air temperature: Development of novel hybrid intelligence model. Energy Rep 7:136\u0026ndash;157. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.egyr.2020.11.033\u003c/span\u003e\u003cspan address=\"10.1016/j.egyr.2020.11.033\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eKumar V, V., Shahin K (2025) Artificial Intelligence and Machine Learning for Sustainable Manufacturing: Current Trends and Future Prospects. Intell Sustainable Manuf 2(1):10002\u0026ndash;10002. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.70322/ism.2025.10002\u003c/span\u003e\u003cspan address=\"10.70322/ism.2025.10002\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWan H, Shen X, Jiang H, Zhang C, Jiang K, Chen T, Shi L, Dong L, He C, Xu Y, Li J, Chen Y (2021) Biomass-derived N/S dual-doped porous hard-carbon as high-capacity anodes for lithium/sodium ions batteries. Energy 231:121102. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.energy.2021.121102\u003c/span\u003e\u003cspan address=\"10.1016/j.energy.2021.121102\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eWeyand A, Rommel C, Zeulner J, Sossenheimer J, Weigold M, Abele E (2021) Method to increase resource efficiency in production with the use of MFCA. Procedia CIRP 98:264\u0026ndash;269. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.procir.2021.01.101\u003c/span\u003e\u003cspan address=\"10.1016/j.procir.2021.01.101\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXia T, An X, Yang H, Jiang Y, Xu Y, Zheng M, Pan E (2023) Efficient Energy Use in Manufacturing Systems\u0026mdash;Modeling, Assessment, and Management Strategy. Energies 16(3):1095. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.3390/en16031095\u003c/span\u003e\u003cspan address=\"10.3390/en16031095\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eXuan Ao H, Eftekhari M, Aungkulanon P, Almutairi K, Mostafaeipour A (2022) Impact of economic and government investment in residential solar power plant on energy system sustainability. Sustain Energy Technol Assess 52:102050. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.seta.2022.102050\u003c/span\u003e\u003cspan address=\"10.1016/j.seta.2022.102050\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang H, Liu S (2022) Sliding Behavior of Silica Ball\u0026ndash;Shale Rock Contact Under Polyacrylamide Aqueous Solutions. J Tribol 144(2):021204. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1115/1.4052682\u003c/span\u003e\u003cspan address=\"10.1115/1.4052682\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003cli\u003e\u003cspan\u003eZhang W, Gu J, Zhou X, Li Y, Wang Y, Xue Y, Liu X (2021) Circulating fluidized bed fly ash based multi-solid wastes road base materials: Hydration characteristics and utilization of SO3 and f -CaO. J Clean Prod 316:128355. \u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://doi.org/10.1016/j.jclepro.2021.128355\u003c/span\u003e\u003cspan address=\"10.1016/j.jclepro.2021.128355\" targettype=\"DOI\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"St. Francis College, Brooklyn, New York, USA","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Artificial intelligence, genetic algorithm, sustainable manufacturing, municipal solid waste, environmental impact, machine learning","lastPublishedDoi":"10.21203/rs.3.rs-7278542/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-7278542/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe paper examines the ways of how artificial intelligence can be \u0026nbsp;\u0026nbsp;utilized in maximizing the sustainability of industry and municipal waste \u0026nbsp;\u0026nbsp;treatment systems. Applying the case of automotive assembly line operations \u0026nbsp;\u0026nbsp;and municipal waste management in Gazipur City Corporation of Bangladesh, we \u0026nbsp;\u0026nbsp;show the potential of AI in resolving resource optimization issues on various \u0026nbsp;\u0026nbsp;levels. In analysis done on automotive assembly line, Genetic Algorithm (GA) \u0026nbsp;\u0026nbsp;optimization was used to reduce energy consumption and waste generation, \u0026nbsp;\u0026nbsp;whereas the municipal waste management analysis showed that there are huge \u0026nbsp;\u0026nbsp;optimization chances in certain cities, such as Gazipur where 1747.121 tons \u0026nbsp;\u0026nbsp;of municipal wastes generate each day with just 60 percent collection \u0026nbsp;\u0026nbsp;efficiency. In an industrial context, GA optimization resulted in a 10 % \u0026nbsp;\u0026nbsp;energy decrease (specifically, ~2200 kWh to ~2000 kWh per run) as well as a \u0026nbsp;\u0026nbsp;25 percent decrease in waste (particularly, ~22 units to ~17 units). These \u0026nbsp;\u0026nbsp;findings show that AI has the potential to help in the ineffective processes \u0026nbsp;\u0026nbsp;in manufacturing as well as managing municipal solid waste. The results \u0026nbsp;\u0026nbsp;indicate that analogous optimization plans ought to enhance waste collection \u0026nbsp;\u0026nbsp;performance in a fast-growing city such as Gazipur City Corporation, which is \u0026nbsp;\u0026nbsp;projected to experience an increase in waste generation up to 5496.80 \u0026nbsp;\u0026nbsp;tons/day by 2035. The findings of this research can serve as evidence-based \u0026nbsp;\u0026nbsp;information that can be offered to the decision-makers to implement AI-based \u0026nbsp;\u0026nbsp;sustainability solutions in industrial and municipal settings.\u003c/p\u003e","manuscriptTitle":"Artificial Intelligence for Clean Industry: A Multi-Agent Framework for Sustainable Waste, Energy, and Emission Practices in Gazipur Factories","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-08-05 14:02:10","doi":"10.21203/rs.3.rs-7278542/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d139ed1d-519c-4664-afb7-1a45f829ba3b","owner":[],"postedDate":"August 5th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":52540647,"name":"Environmental Engineering"}],"tags":[],"updatedAt":"2025-08-05T14:02:10+00:00","versionOfRecord":[],"versionCreatedAt":"2025-08-05 14:02:10","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-7278542","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-7278542","identity":"rs-7278542","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}