Sustainable urban waste collection using a hybrid heuristic–genetic approach: a Bangkok case study

Urban waste collection is a critical component of sustainable city development, directly influencing emissions reduction, resource efficiency, and public health. This study develops a hybrid optimization framework combining a Nearest Neighbor Heuristic with a Genetic Algorithm (GA) to optimize municipal waste collection routes in Bangkok, addressing the Vehicle Routing Problem (VRP) under real-world constraints such as vehicle capacity, time windows, and traffic conditions. The optimized algorithm reduced weekly travel distance by 8.51% and increased average vehicle utilization by 7.78%, translating into projected five-year economic benefits of over 4.7 million Baht and annual GHG emission reduction equivalent to planting approximately 1,750 trees. These findings demonstrate how algorithmic optimization can advance SDG 11 (sustainable cities and communities) and SDG 12 (responsible consumption and production) by aligning technical innovation with environmental and social outcomes. Beyond Bangkok, the framework is scalable to other rapidly urbanizing contexts, offering policymakers a data-driven pathway toward inclusive, low-carbon, and effective waste management systems.

1 Introduction

Bangkok’s waste collection system is increasingly burdened by rising waste volumes and limited municipal capacity (Wannawilai et al., 2017). Effective waste management is essential for advancing sustainable urban development, as it mitigates environmental pollution and supports healthier living conditions. Poorly designed collection routes result in unnecessary fuel consumption, higher operational expenditures, and service inefficiencies (Cardenas et al., 2017; Andruetto et al., 2024). With continued population growth, waste generation intensifies, further straining municipal collection networks (Hajam et al., 2023; Zhang et al., 2024). In the absence of efficient strategies, cities face escalating pollution, public health risks, and rising costs (Abubakar et al., 2022; Alsabt et al., 2024). To counter these challenges and promote long-term sustainability, route optimization has become a prominent focus in urban logistics research (Mancini et al., 2014; Zhao et al., 2024).

Urban waste management has emerged as a critical global concern, particularly in densely populated cities where infrastructure often struggles to keep pace with demand (Voukkali et al., 2024). As urban areas expand, the integration of technology and sustainability becomes increasingly important for effective waste practices. Sustainable cities are progressively adopting advanced computational methods, such as heuristic and genetic algorithms, to enhance efficiency while maintaining ecological balance. However, many municipalities continue to rely on outdated collection systems that are unable to adapt to rapidly changing urban conditions (Morrissey and Browne, 2004). Such inefficiencies exacerbate environmental degradation, increase fuel consumption, and impose avoidable labor expenditures (Omer, 2008; Ekins and Zenghelis, 2021; Magazzino et al., 2022). Algorithmic approaches, including heuristic techniques and Genetic Algorithms (GA), offer a means to optimize vehicle routing, thereby improving collection efficiency, reducing operational costs, and advancing environmental sustainability (Gil et al., 2022; Maroof et al., 2024).

In Asian megacities such as Jakarta, Manila, and Mumbai, municipal waste systems remain heavily burdened by inadequate infrastructure, dense populations, and the persistence of informal waste handling practices (Imura et al., 2005; Aprilia, 2021; Ali et al., 2023; Paul, 2024; Kamalarajan and Swaminathan, 2025). Such systemic constraints aggravate inefficiencies, manifesting in persistent service breakdowns, increased financial burdens, and deteriorating environmental conditions. Contemporary research underscores the need to integrate optimization approaches within broader sustainability frameworks. A mixed-integer optimization algorithm for waste facility planning in Nigeria was developed, through which cost savings and improved recovery outcomes were demonstrated via strategic routing and infrastructure design (Barma et al., 2022). Machine learning was employed to forecast waste generation patterns, allowing more adaptive and efficient collection systems to be achieved (Liu et al., 2025).

The novelty of this study lies in finding route optimization using the combination of nearest-neighbor heuristics and GA with data mining. Using Association Rule Mining and K-Means Clustering, this study identified waste patterns before constructing routes. This creates an optimization algorithm tailored to the real constraints of Asian cities. Beyond technical contributions, this study offers policy recommendations for Bangkok, demonstrating how data-driven optimization can support inclusive, cost-effective, and sustainable urban waste management.

1.1 Literature review

The VRP seeks to identify cost-efficient routes for fleets serving multiple destinations. Initially formulated by Clark and Wright in 1964 to address logistics challenges (Clark and Wright, 1964), the VRP has since expanded to incorporate constraints such as time windows and vehicle capacities (Laporte, 1992). Its applications extend across transportation logistics, emergency response, and waste collection, where optimization contributes not only to efficiency but also to reduced fuel use and lower emissions. Variants such as the Capacitated VRP (CVRP) and VRP with stochastic demands further demonstrate the adaptability of the model to complex, sustainability-oriented urban systems (Gendreau et al., 1996).

VRPTW extends the classical VRP by incorporating delivery time constraints. Research on route construction, local search algorithms, and metaheuristics has demonstrated the effectiveness of these approaches in solving VRPTW challenges in urban logistics (Bräysy and Gendreau, 2005a,b). In the context of waste collection, time-window integration ensures reliable service delivery while simultaneously optimizing vehicle capacity and travel efficiency, thereby reducing unnecessary fuel consumption. VRPTW applications have become central to smart logistics and sustainable waste management systems (Desaulniers et al., 2014).

Heuristic methods such as Nearest Neighbor and the Clark-Wright Savings algorithm remain widely applied in solving VRP (Garside et al., 2024). These approaches generate rapid, near-optimal solutions, making them particularly suitable for operational contexts where timely decisions are critical. Their flexibility also supports dynamic routing scenarios that require real-time adjustments, a feature essential for sustainable urban services. More recently, hybrid heuristic approaches have been developed that integrate machine learning, enabling improved decision-making and adaptive optimization in waste collection and other sustainability-driven applications (Dieter et al., 2023).

GAs, inspired by evolutionary principles (Holland, 1992), have been extensively applied to VRP (Baker and Ayechew, 2003). By iteratively refining route assignments through selection, crossover, and mutation, GAs generate high-quality solutions for complex routing problems. Their effectiveness has been demonstrated in large-scale logistics and urban infrastructure planning, where efficiency gains also translate into reduced energy use and environmental impacts (Xia et al., 2023; Tang, 2025). Recent advancements, including adaptive mutation and crossover strategies, have further enhanced GA convergence, making them increasingly relevant for sustainable urban waste collection optimization (Zainuddin and Samad, 2020).

Although VRP and its variants have been widely studied, important gaps remain in sustainable urban waste collection. In rapidly urbanizing Asian contexts, where traffic and waste dynamics present unique challenges, research tends to apply heuristics or genetic algorithms separately, rather than combining them in hybrid approaches that could strengthen both efficiency and sustainability outcomes. Moreover, the integration of data mining with route optimization to anticipate waste generation patterns is underdeveloped, despite its potential for adaptive and resource-efficient service delivery. Finally, cluster-based optimization methods that capture spatial and temporal variation in waste flows are scarce, particularly in dense urban contexts such as Bangkok, where sustainability goals intersect with equity and operational efficiency.

1.2 Contributions

Building upon existing research, this study advances sustainable waste collection optimization through the following contributions.

1) It develops a hybrid optimization framework that combines heuristic methods with a GA, addressing the limitations of using either approach in isolation, and improving route efficiency in dense urban contexts.

2) It applies data-driven techniques, specifically association rule mining and K-means clustering, to identify waste generation patterns that inform dynamic and adaptive route planning.

3) It demonstrates the integration of hybrid optimization with data-driven insights in the context of Bangkok, providing an evidence-based framework that can be adapted to other rapidly urbanizing cities facing similar sustainability challenges.

4) It offers preliminary policy insights, drawing on global practices and the Bangkok case study, to highlight how supportive measures such as government incentives, regulatory requirements, Public-Private Partnerships (PPPs), and economic instruments could enhance the adoption and scalability of optimized waste collection systems.

The remainder of this paper is organized as follows. Section 1 introduces the study, reviews the relevant literature, and outlines the contributions. Section 2 details the research methodology. Section 3 reports the results of the optimization algorithm. Section 4 provides a discussion of the findings and their broader implications. Section 5 concludes with key insights and highlights relevant points for further study.

2 Methodology

This study employs heuristic methods and GA to optimize waste collection routes in Lat Krabang District, Bangkok. The methodology consists of three key steps: data collection and route mapping, algorithm implementation (see Figure 1), and algorithm evaluation. The algorithm implementation consists of development of a heuristic algorithm, and optimization using a genetic algorithm.

Figure 1

Figure 1. Methodological flowchart.

In this study, each waste collection point must be serviced within a specified time window determined by municipal regulations. Specifically, collection trucks can only operate between 05:00 and 20:00, and certain neighborhoods have additional service-time constraints due to local ordinances. Additionally, each vehicle has a maximum load capacity of 6,500 kg, which limits the number of collection points that can be serviced on a single route. These constraints shape route planning to ensure timely collection, minimize traffic disruptions, comply with municipal guidelines, and maximize operational efficiency.

2.1 Data collection and route mapping

The study collected data on waste collection routes across Lat Krabang District in four main categories:

1) Geographical data were gathered by recording GPS coordinates of waste collection points and depot locations through site visits. This ensured accuracy in route planning.

2) Operational data were obtained by measuring daily waste generation volumes. Trucks were weighed at the depot before departure and after collection, and the net values were recorded (see Table 1. These measurements were essential for vehicle capacity planning and scheduling.

3) Existing route data were compiled from municipal records, which provided current schedules and distances. These data served as a baseline for optimization and revealed inefficiencies in the existing system.

4) Traffic considerations were incorporated by analyzing congestion patterns. Municipal traffic density records were used and validated against Google Maps congestion indices during peak hours to create realistic route schedules and avoid delays.

Table 1

Table 1. Waste collection routes and waste volume.

Data collection was conducted over a two-week period through on-site surveys. In total, 244,340.80 kg of municipal solid waste were recorded across seven daily routes and 91 vehicles. This corresponds to an average daily waste volume of approximately 34,905.83 kg. These inputs formed the basis of the VRP algorithm, ensuring that the optimization accurately reflected real-world operational constraints.

2.2 Algorithm implementation

The proposed optimization process algorithm followed a two-phase approach that combined heuristic preprocessing with GA.

2.2.1 Heuristic implementation

An initial route assignment was generated using a nearest–neighbor heuristic as shown in Algorithm 1. In this procedure, an empty route was created at the depot, and the closest unvisited waste collection point was iteratively added if the truck still had capacity. Once the truck was filled, it returned to the depot, and the process repeated until all points were serviced. This approach produced a feasible, though not necessarily optimal, initial solution for each day’s routes.

ALGORITHM 1

ALGORITHM 1. Nearest Neighbour Heuristic in Vehicle Routing Problem (NNH-VRP)

The nearest-neighbor heuristic involving three main steps:

1) The algorithm identified the nearest unvisited waste collection point by calculating the distance from the vehicle’s current location to all remaining unvisited points and selecting the closest one to minimize travel time and fuel consumption.

2) The selected collection point was then assigned to the nearest available vehicle. When multiple vehicles were available, the algorithm determined which vehicle could reach the collection point most efficiently.

3) The process continued until all waste points were assigned, while ensuring that vehicle capacity constraints were respected. This preprocessing step provided an initial feasible solution that was subsequently refined using GA.

2.2.2 Genetic algorithm implementation

As detailed in Algorithm 2, the GA optimization process was structured around six key components. First, chromosome representation encoded routes as ordered sequences of waste collection points. Second, the fitness function was designed to minimize total route distance while balancing vehicle loads. Third, tournament selection was applied to enhance convergence. Fourth, order-based crossover was used to preserve feasible route sequences. Fifth, swap mutation was introduced to maintain diversity and prevent premature convergence to local optima. Finally, stopping criteria were defined so that the algorithm terminated when no significant improvement was observed over multiple iterations.

ALGORITHM 2

ALGORITHM 2. Genetic Algorithm for Nearest Neighbour Heuristic in Vehicle Routing Problem (GA-NNH-VRP)

The optimized parameters for GA are summarized in Table 2. A mutation probability of 0.2 and a crossover probability of 0.9 were selected after preliminary trials, which indicated faster convergence without premature stagnation. In most cases, the algorithm converged within 6,000–8,000 generations, after which improvements in route distance leveled off. An adaptive mechanism was also incorporated, slightly increasing the mutation rate if the best solution failed to improve for 200 consecutive generations. This adjustment helped the search escape local optima and improved robustness.

Table 2

Table 2. Parameters for GA.

2.3 Algorithm evaluation

The proposed algorithm was evaluated using three performance metrics that reflect operational efficiency and sustainability outcomes.

1) Total route distance reduction was calculated to quantify how much the total distance traveled by all waste collection vehicles decreased after applying the optimized routes compared with the original routes.

2) Vehicle utilization efficiency was assessed to determine how effectively the vehicles’ capacity was used. This metric was measured as the percentage of the vehicle’s volume or weight capacity that was filled with waste on average.

3) Operational cost savings were estimated to capture the overall financial benefits of implementing the optimized routes. This measure reflected reductions in fuel consumption, maintenance costs, and other distance-related expenses.

3 Results

This section presents the results obtained from implementing the heuristic and genetic algorithm-based waste collection optimization algorithm. The data was analyzed by comparing the optimized routes with the existing routes based on total distance travel, vehicle utilization, and operational cost efficiency.

3.1 Route distance and waste collection optimization

The proposed algorithm was tested on waste collection routes for each day of the week (Figure 2). The results are summarized in Table 3, which shows the total waste collected and the reduction in travel distance achieved using the optimized routing algorithm. Also, the summarized Table 3 is graphically presented in a scatter plot to visually further support data interpretation as illustrated in Figure 3.

Figure 2

Figure 2. Example of truck route optimization over a 7-day period: (A) Monday, (B) Tuesday, (C) Wednesday, (D) Thursday, (E) Saturday, and (F) Sunday. Pickup locations are marked with symbols, each color-coded according to the specific day.

Table 3

Table 3. Summary of waste collection optimization results.

Figure 3

Figure 3. Comparison of travel distance before and after optimizations.

The current system assigns fixed routes based on geographical zones, historical waste volumes, community-specific schedules, and operational constraints. However, this does not adjust to daily variations in waste volume, traffic, or urban activity. As a result, collection distances fluctuate with zone waste loads, point distribution, and peak periods in high-density areas. This manual and static route planning approach leads to inefficiencies, including route overlaps and underutilized truck capacity, justifying the need for the hybrid optimization framework developed in this study.

As can be observed in Table 3, the optimized waste collection routes resulted in a significant reduction in travel distance, with an overall reduction of 199.44 km per week (8.51%). The highest reduction in distance was achieved on Monday (14.97%), while the lowest reduction was observed on Saturday (6.87%). This suggests that certain days had more inefficient original routes, which benefited more from the optimization process. Additionally, the optimized algorithm efficiently redistributed workload, preventing route overlaps and unnecessary travel.

Figure 3 illustrates the comparative travel distances for waste collection routes before and after optimization across each day of the week. The plotted data clearly demonstrates a consistent reduction in travel distance following the implementation of the optimized routing algorithm. The most pronounced improvement is observed on Monday, aligning with the highest percentage reduction (14.97%) reported in Table 3. Conversely, Saturday shows the least reduction (6.87%), indicating relatively more efficient original routing on that day. The visual trend supports the conclusion that the hybrid optimization framework effectively minimizes redundant travel and improves route efficiency, with varying degrees of impact depending on the initial route structure and daily waste distribution patterns.

3.2 Vehicle utilization and Fleet efficiency

To further assess the efficiency of the optimized routes, vehicle utilization was compared before and after applying the heuristic-genetic algorithm approach to assess the optimized routes efficiency (see Table 4). The results indicate an average increase of 7.78% in vehicle utilization under the optimized algorithm. Notably, Saturday exhibited the highest improvement at 17.02%, suggesting a significant rebalancing of workload across the fleet. In contrast, Wednesday and Sunday showed no change, implying that route distribution on those days was already relatively balanced prior to optimization.

Table 4

Table 4. Vehicle utilization before and after optimization.

Figure 4 visually supports these results, showing a general upward shift in utilization after optimization. This graphical evidence highlights improved fleet efficiency and balanced operations, aligning with the findings in Table 4 and underscoring the effectiveness of the heuristic–genetic algorithm approach.

Figure 4

Figure 4. Percentage of vehicle utilization comparison before and after optimizations.

These improvements reflect more effective use of vehicle capacity, minimizing instances of underutilized collection trucks. By enhancing load distribution and route efficiency, the optimization algorithm enables each truck to carry a greater volume of waste per trip. This, in turn, has the potential to reduce the total number of vehicles required to perform the same collection workload, thereby lowering operational costs and environmental impact.

For the Wednesday and Sunday routes, the optimization algorithm did not produce significant changes. This outcome may reflect that the existing routes were already operating near an efficient state under the current constraints, such as fixed collection points, vehicle capacities, or time limits.

According to the principle of diminishing marginal returns in combinatorial optimization, solutions that are already close to a local optimum tend to show limited further improvement, even when advanced algorithms are applied.

Therefore, the lack of improvement on these days does not suggest that the algorithm is unnecessary. This indicates that the algorithm can identify near-optimal configurations and recognizes when further optimization is constrained by existing operational limits.

3.3 Economic and environmental impact analysis

The reduction in travel distance directly contributes to lower fuel consumption and Greenhouse Gas (GHG) emissions. Optimized routes lead to fewer kilometers traveled, which translates to decreased fuel usage. This positively impacts both economic and environmental aspects:

1) Fuel savings: Reduced fuel consumption leads to lower operational costs.

2) Lower GHG emissions: A decrease in total travel distance reduces GHG emissions.

3) Improved service efficiency: More evenly distributed vehicle loads result in consistent service levels across all routes.

4) Reduced Maintenance Costs: Fewer kilometers of travel result in reduced vehicle deterioration and depreciation, thereby extending the lifespan of the vehicles.

4 Discussion

The results confirm that combining heuristic methods with GA effectively reduces travel distance and improves vehicle utilization. The optimization algorithm demonstrated robust performance across multiple days, indicating its ability to handle various waste collection constraints and urban traffic conditions.

The findings are consistent with previous studies advocating the use of metaheuristic optimization techniques in municipal waste management. Compared to traditional routing methods, the hybrid approach achieves:

1) Higher efficiency in route optimization: A more balanced distribution of waste collection points resulted in shorter routes and reduced bottlenecks.

2) Scalability to larger urban areas: The adaptability of the genetic algorithm suggests that the algorithm could be implemented in other cities with different waste collection requirements.

The proposed algorithm is sensitive to operational data such as daily waste generation volumes, GPS-defined collection points, working hours, and available vehicle capacity. To ensure sustained efficiency, it is recommended that route optimization algorithms be periodically updated using the latest data. Future work should conduct a formal sensitivity analysis to assess how fluctuations in these parameters influence route selection, vehicle utilization, and cost projections.

4.1 Extended analysis

4.1.1 Data mining and clustering analysis

Such analyses for this exploratory demonstrate how data mining and clustering can complement optimization algorithms by revealing temporal, spatial, and economic patterns that support sustainable waste management strategies.

1) Association rule for mining

Association rule mining using the Apriori method was applied to identify patterns between waste generation volumes and collection areas, providing complementary insights for route planning (see Table 5). The analysis revealed that high-traffic areas were strongly associated with higher waste volumes (confidence = 0.79), supporting the need for targeted capacity planning in these zones. Commercial districts showed a significant correlation with weekend peaks (confidence = 0.83), indicating the importance of allocating additional resources during weekends. Residential areas exhibited Monday peaks (confidence = 0.76), likely reflecting accumulated weekend waste, while educational institutions such as King Mongkut’s Institute of Technology Ladkrabang (KMITL) demonstrated weekday consistency (confidence = 0.94), suggesting predictable collection needs. These findings provide actionable evidence for dynamic route adjustments: routes serving high-traffic areas, such as Routes 3 and 6, should receive priority allocation during peak hours, whereas Route 7 (KMITL) benefits from standardized scheduling due to its stable waste generation pattern.

Table 5

Table 5. Association rules for waste generation patterns and collection planning.

2) K-Means clustering

K-Means clustering was applied to group collection points by waste volume, geographical proximity, and temporal patterns, with the elbow method identifying four optimal clusters that informed route optimization outcomes (see Table 6). Cluster 1, representing high-volume areas, corresponded primarily to Route 3 and explained the significant Monday distance reduction (14.97%), as dense waste loads enabled more efficient vehicle utilization (Table 4). Cluster 2, comprising medium-volume residential areas, showed strong day-of-week variation, with Monday collections 31.5% above average, accounting for the pronounced Monday optimization gains. Cluster 3, covering low-volume and sparse areas such as Route 7 (KMITL), achieved the highest potential for distance reduction (18.63%) by minimizing deadhead trips, while its consistent weekday pattern supported improved Saturday utilization (17.02%). Cluster 4, consisting of variable-volume commercial areas, exhibited high day-to-day variability (coefficient of variation = 0.37), aligning with moderate improvements in Route 2. Overall, these cluster-specific patterns clarify why optimization effects varied across days and routes, reinforcing the algorithm’s ability to adapt to heterogeneous urban waste generation dynamics.

Table 6

Table 6. Cluster characteristics and their impact on optimization.

3) Economic and environmental impact analysis

The extended economic analysis projects substantial cost savings and environmental benefits from implementing the optimized routing system (see Table 7). Projections were based on current operational costs and observed performance improvements, with five-year estimates incorporating a 3% annual inflation rate for fuel and labor, 2% annual growth in waste volume, and a conservative 5% discount rate to calculate the net current value of future savings. Implementation costs—including software development (320,000 Baht), staff training (180,000 Baht), and annual system maintenance (120,000 Baht)—were included in the calculation of a 14-month breakeven point.

Table 7

Table 7. Cost savings and environmental benefits.

The results indicate a total projected five-year economic benefit of 4,742,910 Baht, representing a significant Return On Investment (ROI). The environmental analysis further demonstrates that optimized routing contributes directly to Bangkok’s GHG emission reduction goals, with projected annual emission savings equivalent to planting approximately 1,750 trees. Emission reduction was calculated using the standard diesel emission factor of 2.74062321 kg CO₂ eq./L [Thailand Greenhouse Gas Management Organization (Public Organization), 2022].

These findings align with sustainability principles by simultaneously reducing GHG emissions and improving cost efficiency in municipal operations. When integrated with policy frameworks that emphasize environmental conservation and social inclusivity, the adoption of such optimization algorithms can amplify both economic and environmental benefits for urban communities.

4.2 Policy framework

Drawing from the best global practices and the results of the Bangkok case study, this framework integrates four core pillars to achieve sustainable waste collection system as follows:

4.2.1 Government incentives

The implementation of sustainable waste management requires strong government support through various incentive mechanisms. Policy incentives are necessary to encourage adoption in developing Asian cities (Wilkerson et al., 2022; Brimblecombe et al., 2023). Strategic infrastructure development, including optimized facility placement and transportation routes, has shown potential cost reductions of up to 23% in waste collection expenses (Hemidat et al., 2017). These incentives directly support the efficiency gains demonstrated in this study, where optimized routing reduced weekly travel distance by 8.51% and generated measurable fuel and emission saving, indicating that government-backed adoption can scale these environmental and economic benefits.

4.2.2 Mandatory requirements

Regulatory requirements play a crucial role in ensuring effective waste management practices. Municipalities should adopt hybrid algorithms combining heuristics and GA to reduce travel distances and improve vehicle utilization (Kamsopa et al., 2021). These algorithms balance computational efficiency and solution quality, ensuring routes adapt to dynamic urban conditions like traffic congestion and variable waste volumes.

In addition, enforce GPS tracking to monitor fleet movements and dynamically adjust routes. Systems use IoT sensors to optimize collection schedules, reducing fuel costs and enabling real-time adjustments during peak traffic (Moustakas, 2025). This study demonstrates that such mandatory efficiency standards are justified by quantifiable results: vehicle utilization improved by 7.78% (with Saturday showing a 17.02% improvement), and 199.44 km of unnecessary weekly travel were eliminated. Enforcing GPS tracking and hybrid routing standards would help institutionalize these efficiencies across municipal operators.

4.2.3 Public-private partnerships (PPPs)

PPPs should be adopted as a central policy instrument for advancing sustainable solid waste management systems (SBN Software, 2025; Chatri and Aziz, 2012; Madinah, 2016). Municipalities can leverage private-sector expertise, technology, and investment to enhance efficiency, scalability, and innovation, while distributing financial and operational risks between stakeholders (SBN Software, 2025; Obani et al., 2025; Yeboah, 2024).

To ensure effectiveness, PPP frameworks must be guided by clear policy principles. First, regulatory clarity is essential: transparent bidding and contracting processes, well-defined roles and responsibilities, and compliance with environmental and public health standards must be mandated (Chatri and Aziz, 2012; Madinah, 2016; Yeboah, 2024; Trafford and Proctor, 2006; da Cruz et al., 2013). Second, accountability and oversight should be embedded through measurable performance indicators and monitoring mechanisms. Third, inclusive stakeholder engagement is required: participatory planning processes must involve local communities and informal sector actors to ensure equity, build trust, and strengthen adoption. Finally, adaptive governance should be institutionalized so that PPPs remain responsive to evolving urban conditions, technological advances, and sustainability targets.

By embedding these principles into policy, PPPs can become a reliable mechanism for mobilizing resources, accelerating digital innovation, and achieving long-term sustainability goals in municipal waste management. In this study, the Lat Krabang optimization model projects a five-year economic benefit of approximately 4.7 million Baht, primarily from fuel and maintenance savings; a performance-based PPP model would allow the municipality and private operators to share these savings, creating direct incentives to maintain and continuously improve the optimized routing system.

4.2.4 Economic incentives

Financial mechanisms drive the adoption of optimized routing systems by aligning economic incentives with sustainability goals. Pay-As-You-Throw (PAYT) (Messina et al., 2023) systems charge households/businesses based on waste generation, incentivizing reduction and recycling while generating revenue to fund route optimization software. Subsidies and tax breaks support municipalities adopting electric vehicles, leveraging projected annual fuel savings to offset upfront costs, alongside tax rebates for achieving recycling targets. In addition, green bonds redirect savings from reduced travel distances into financing recycling infrastructure or solar-powered waste stations, while revolving funds sustain innovation in AI-driven predictive modeling and cluster-based scheduling. These instruments collectively amplify the study’s demonstrated benefits, including annual GHG emissions reductions and improved vehicle utilization, creating a self-reinforcing cycle of economic efficiency and environmental stewardship (Morlok et al., 2017).

The five-year economic benefit of 4.7 million Baht and the associated annual savings in fuel, maintenance, labor, and GHG emissions in this study demonstrate that economic instruments can sustain the financial and environmental gains from optimized routing.

4.3 Limitations and future directions

While the study demonstrates the effectiveness of combining the heuristic method with GA for municipal waste collection, several limitations should be acknowledged. First, the dataset was limited to a single district and a two-week observation period, which may not fully capture seasonal or long-term variations in waste generation. These variations could affect the optimized routes which potentially requiring dynamic adjustments to fleet size or frequency that this study did not predict. Second, because this study is based on Lat Krabang District, which may differ from other districts in road layout and traffic patterns. As a result, key parameters would need to be adjusted before applying the model to other urban areas. Third, the optimization model incorporated static traffic assumptions and indicative operational parameters; real-time traffic dynamics, driver behavior, and unexpected disruptions were not modeled. Third, the economic projections, while comprehensive, relied on assumed inflation rates, growth trends, and discount factors that may vary under different economic conditions.

Future research should extend the analysis across multiple districts and longer timeframes to validate generalizability. Integrating IoT sensors, GPS tracking, and real-time traffic data would enhance the adaptability of routing algorithms under dynamic urban conditions. Comparative studies across cities with different socio-economic and infrastructural contexts could further test transferability. Finally, incorporating stakeholder perspectives, including municipal authorities, private operators, and local communities, would strengthen the policy relevance and practical adoption of optimization frameworks.

5 Conclusion

This study developed and applied a hybrid heuristic–genetic algorithm to optimize municipal waste collection in Bangkok, demonstrating measurable improvements in route efficiency, vehicle utilization, cost savings and lower emissions. By integrating data mining and clustering analyses, the algorithm also revealed temporal and spatial waste generation patterns that support adaptive scheduling. The combined technical and analytical approach reduced weekly travel distance by 8.51%, increased average vehicle utilization by 7.78%, and projected a five-year economic benefit exceeding 4.7 million Baht, alongside annual emission reduction equivalent to planting 1,750 trees.

By directly reducing urban transport emissions and improving the resilience and inclusivity of municipal waste services, the study demonstrates how algorithmic optimization can advance multiple sustainability targets simultaneously.

Beyond operational gains, the findings underscore the importance of embedding optimization within broader sustainability frameworks. Policy instruments such as government incentives, regulatory requirements, PPPs, and economic mechanisms can accelerate adoption and ensure equitable, long-term benefits. While the study was limited to a single district and short observation period, the framework is transferable to other urban contexts, provided that local data and governance structures are integrated.

Future research should extend validation across multiple districts and longer timeframes, incorporate real-time traffic and IoT-enabled monitoring, and explore multi-objective optimization that balances efficiency, equity, and environmental outcomes. By aligning algorithmic innovation with policy support, municipalities can advance toward more resilient, low GHG emissions, and inclusive waste management systems that directly contribute to SDG 11 (sustainable cities and communities) and SDG 12 (responsible consumption and production).

Data availability statement

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

Author contributions

CH: Conceptualization, Methodology, Resources, Software, Supervision, Writing – original draft. JK: Writing – original draft, Writing – review & editing, Conceptualization, Funding acquisition, Project administration, Validation. AS: Writing – original draft, Writing – review & editing, Data curation, Formal analysis, Investigation, Methodology, Visualization.

Funding

The author(s) declared that financial support was received for this work and/or its publication. This work was financially supported by School of Engineering, King Mongkut’s Institute of Technology Ladkrabang under Grant no: 2564-02-01-074.

Acknowledgments

The authors would like to thank King Mongkut’s Institute of Technology Ladkrabang for providing the necessary facilities, and Lat Krabang District, Bangkok for its assistance.

Conflict of interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Generative AI statement

The author(s) declared that Generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

References

Abubakar, I. R., Maniruzzaman, K. M., Dano, U. L., AlShihri, F. S., AlShammari, M. S., Ahmed, S. M. S., et al. (2022). Environmental sustainability impacts of solid waste management practices in the global south. Int. J. Environ. Res. Public Health 19:12717. doi: 10.3390/ijerph191912717,

PubMed Abstract | Crossref Full Text | Google Scholar

Ali, A. A., Golbert, Y., Reksa, A. F. A., Kretzer, M. M., and Schweiger, S. (2023). “Transformative solutions in the global south: addressing solid waste management challenges in Jakarta through participation by civil society organizations?” in Environmental governance in Indonesia. eds. A. Triyanti, M. Indrawan, L. Nurhidayah, and M. A. Marfai (Cham: Springer), 329–351.

Google Scholar

Alsabt, R., Alkhaldi, W., Adenle, Y. A., and Alshuwaikhat, H. M. (2024). Optimizing waste management strategies through artificial intelligence and machine learning – an economic and environmental impact study. Clean. Waste Syst. 8:100158. doi: 10.1016/j.clwas.2024.100158

Crossref Full Text | Google Scholar

Andruetto, C., Stenemo, E., and Pernestål, A. (2024). Towards sustainable urban logistics: exploring the implementation of city hubs through system dynamics. Transp. Res. Interdiscip. Perspect. 27:101204. doi: 10.1016/j.trip.2024.101204

Crossref Full Text | Google Scholar

Aprilia, A. (2021). Waste management in Indonesia and Jakarta: challenges and way forward. Singapore: Asia-Europe Foundation (ASEF).

Google Scholar

Baker, B. M., and Ayechew, M. A. (2003). A genetic algorithm for the vehicle routing problem. Comput. Oper. Res. 30, 787–800. doi: 10.1016/S0305-0548(02)00051-5

Crossref Full Text | Google Scholar

Barma, M., Biniyamin, H. K., Modibbo, U. M., and Gaya, H. M.'a. (2022). Mathematical model for the optimization of municipal solid waste management. Front. Sustain. 3. doi: 10.3389/frsus.2022.880409

Crossref Full Text | Google Scholar

Bräysy, O., and Gendreau, M. (2005a). Vehicle routing problem with time windows, part I: route construction and local search algorithms. Transp. Sci. 39, 104–118. doi: 10.1287/trsc.1030.0056,

PubMed Abstract | Crossref Full Text | Google Scholar

Bräysy, O., and Gendreau, M. (2005b). Vehicle routing problem with time windows, part II: metaheuristics. Transp. Sci. 39, 119–139. doi: 10.1287/trsc.1030.0057,

PubMed Abstract | Crossref Full Text | Google Scholar

Brimblecombe, J., Miles, B., Chappell, E., De Silva, K., Ferguson, M., Mah, C., et al. (2023). Implementation of a food retail intervention to reduce purchase of unhealthy food and beverages in remote Australia: mixed-method evaluation using the consolidated framework for implementation research. Int. J. Behav. Nutr. Phys. Act. 20:20. doi: 10.1186/s12966-022-01377-y,

PubMed Abstract | Crossref Full Text | Google Scholar

Cardenas, I., Borbon-Galves, Y., Verlinden, T., de Van Voor, E., Vanelslander, T., and Dewulf, W. (2017). City logistics, urban goods distribution and last mile delivery and collection. Compet. Regul. Netw. Ind. 18, 22–43.

Google Scholar

Chatri, A. K., and Aziz, A. (2012). Public private partnerships in solid waste management: potential and strategies. Chennai: Athena Infonomics India.

Google Scholar

Clark, G., and Wright, J. W. (1964). Scheduling of vehicles from a central depot to a number of delivery points. Oper. Res. 12, 568–581. doi: 10.1287/opre.12.4.568,

PubMed Abstract | Crossref Full Text | Google Scholar

da Cruz, N. F., Simões, P., and Marques, R. C. (2013). The hurdles of local governments with PPP contracts in the waste sector. Environ. Plan. C Gov. Policy 31, 292–307. doi: 10.1068/c11158

Crossref Full Text | Google Scholar

Desaulniers, G., Madsen, O. B. G., and Ropke, S. (2014). “The vehicle routing problem with time windows” in Vehicle routing: problems, methods, and applications. eds. P. Toth and D. Vigo (Philadelphia, PA: Society for Industrial and Applied Mathematics (SIAM)), 119–159.

Google Scholar

Dieter, P., Caron, M., and Schryen, G. (2023). Integrating driver behavior into last-mile delivery routing: combining machine learning and optimization in a hybrid decision support framework. Eur. J. Oper. Res. 311, 283–300. doi: 10.1016/j.ejor.2023.04.043

Crossref Full Text | Google Scholar

Ekins, P., and Zenghelis, D. (2021). The costs and benefits of environmental sustainability. Sustain. Sci. 16, 949–965. doi: 10.1007/s11625-021-00910-5,

PubMed Abstract | Crossref Full Text | Google Scholar

Garside, A. K., Ahmad, R., and Muhtazaruddin, M. N. B. (2024). A recent review of solution approaches for green vehicle routing problem and its variants. Oper. Res. Perspect. 12:100303. doi: 10.1016/j.orp.2024.100303

Crossref Full Text | Google Scholar

Gendreau, M., Laporte, G., and Semet, F. (1996). A Tabu search heuristic for the VRP with stochastic demands and customers. Oper. Res. 44, 469–477. doi: 10.1287/opre.44.3.469,

PubMed Abstract | Crossref Full Text | Google Scholar

Gil, A. F., Lalla-Ruiz, E., Sánchez, M. G., and Castro, C. (2022). A review of heuristics and hybrid methods for green vehicle routing problems considering emissions. J. Adv. Transp. 2022, 1–38. doi: 10.1155/2022/5714991,

PubMed Abstract | Crossref Full Text | Google Scholar

Hajam, Y. A., Kumar, R., and Kumar, A. (2023). Environmental waste management strategies and vermi transformation for sustainable development. Environ. Chall. 13:100747. doi: 10.1016/j.envc.2023.100747

Crossref Full Text | Google Scholar

Hemidat, S., Oelgemöller, D., Nassour, A., and Nelles, M. (2017). Evaluation of key indicators of waste collection using GIS techniques as a planning and control tool for route optimization. Waste Biomass Valoriz. 8, 1533–1554. doi: 10.1007/s12649-017-9938-5

Crossref Full Text | Google Scholar

Holland, J. H. (1992). Adaptation in natural and artificial systems: an introductory analysis with applications to biology, control, and artificial intelligence. Cambridge, MA: MIT Press.

Google Scholar

Imura, H., Yedla, S., Shirakawa, H., and Memon, M. A. (2005). Urban environmental issues and trends in Asia—an overview. Int. Rev. Environ. Strateg. 5, 357–382.

Google Scholar

Kamalarajan, V., and Swaminathan, L. (2025). Household waste collection and management in urban low-income areas: insights from Mumbai's central suburbs. J. Inform. Educ. Res. 5.

Google Scholar

Kamsopa, K., Sethanan, K., Jamrus, T., and Czwajda, L. (2021). Hybrid genetic algorithm for multi-period vehicle routing problem with mixed pickup and delivery with time window, heterogeneous fleet, duration time and rest area. Eng. J. 25, 71–86. doi: 10.4186/ej.2021.25.10.71

Crossref Full Text | Google Scholar

Laporte, G. (1992). The vehicle routing problem: an overview of exact and approximate algorithms. Eur. J. Oper. Res. 59, 345–358. doi: 10.1016/0377-2217(92)90192-C

Crossref Full Text | Google Scholar

Liu, X., Zhi, W., and Akhundzada, A. (2025). Enhancing performance prediction of municipal solid waste generation: a strategic management. Front. Environ. Sci. 13. doi: 10.3389/fenvs.2025.1553121

Crossref Full Text | Google Scholar

Madinah, N. (2016). Solid waste management system: public-private partnership, the best system for developing countries. Int. J. Eng. Res. Appl. 6, 57–67.

Google Scholar

Magazzino, C., Toma, P., Fusco, G., Valente, D., and Petrosillo, I. (2022). Renewable energy consumption, environmental degradation and economic growth: the greener the richer? Ecol. Indic. 139:108912.

Google Scholar

Mancini, S., Gonzalez-Feliu, J., and Crainic, T. G. (2014). “Planning and optimization methods for advanced urban logistics systems at tactical level” in Sustainable urban logistics: concepts, methods and information systems. eds. J. Gonzalez-Feliu, F. Semet, and J. L. Routhier, vol. xx (Heidelberg: Springer), 145–164.

Google Scholar

Maroof, A., Ayvaz, B., and Naeem, K. (2024). Logistics optimization using hybrid genetic algorithm (HGA): a solution to the vehicle routing problem with time windows (VRPTW). IEEE Access 12, 36974–36989. doi: 10.1109/ACCESS.2024.3373699

Crossref Full Text | Google Scholar

Messina, G., Tomasi, A., Ivaldi, G., and Vidoli, F. (2023). ‘Pay as you own’ or ‘pay as you throw’? A counterfactual evaluation of alternative financing schemes for waste services. J. Clean. Prod. 412:137363. doi: 10.1016/j.jclepro.2023.137363

Crossref Full Text | Google Scholar

Morlok, J., Schoenberger, H., Styles, D., Galvez-Martos, J.-L., and Zeschmar-Lahl, B. (2017). The impact of pay-as-you-throw schemes on municipal solid waste management: the exemplar case of the county of Aschaffenburg, Germany. Resources 6:8. doi: 10.3390/resources6010008

Crossref Full Text | Google Scholar

Morrissey, A. J., and Browne, J. (2004). Waste management models and their applications to sustainable waste management. Waste Manag. 24, 297–308.

Google Scholar

Moustakas, A. (2025). Available online at: https://www.shiftfleet.ai/waste-management-with-gps-fleet-management-telematics/ (Accessed January 20, 2025).

Google Scholar

Obani, I. P., Obani, Z. I., Anaeto, F. C., and Akroh, T. O. (2025). Public-private collaborations in waste management: evaluating policy effectiveness and governance models in Nigeria. J. Integr. Ecosyst. Environ. 3, 25–53.

Google Scholar

Omer, A. M. (2008). Energy, environment and sustainable development. Renew. Sust. Energ. Rev. 12, 2265–2300. doi: 10.1016/j.rser.2007.05.001

Crossref Full Text | Google Scholar

Paul, R. (2024). Manila: waste management. Available online at: https://www.ruchirapaul.com/writing/manila-waste-managment (Accessed August 10, 2025).

Google Scholar

SBN Software. (2025). Available online at: https://sbnsoftware.com/blog/what-role-do-public-private-partnerships-play-in-waste-management/ (Accessed February 20, 2025).

Google Scholar

Tang, W. (2025). Optimization of urban terminal delivery routes using fuzzy genetic algorithm and its practical application. J. Comput. Methods Sci. Eng. 25. doi: 10.1177/14727978251323125

Crossref Full Text | Google Scholar

Thailand Greenhouse Gas Management Organization (Public Organization). (2022). Emission factor (CFO). Available online at: https://thaicarbonlabel.tgo.or.th/index.php?lang=TH&mod=YjNKbllXNXBlbUYwYVc5dVgyVnRhWE56YVc5dQ&utm_source=chatgpt.com (Accessed December 2, 2025).

Google Scholar

Trafford, S., and Proctor, T. (2006). Successful joint venture partnerships: public-private partnerships. Int. J. Public Sect. Manag. 19, 117–129. doi: 10.1108/09513550610650392

Crossref Full Text | Google Scholar

Voukkali, I., Papamichael, I., Loizia, P., and Zorpas, A. A. (2024). Urbanization and solid waste production: prospects and challenges. Environ. Sci. Pollut. Res. 31, 17678–17689. doi: 10.1007/s11356-023-27670-2,

PubMed Abstract | Crossref Full Text | Google Scholar

Wannawilai, P., Poboon, C., and Maneein, J. (2017). Analysis of solid waste management and strategies for Bangkok metropolitan. Environ. Nat. Resour. J. 15, 1–12.

Google Scholar

Wilkerson, B., Romanenko, E., and Barton, D. N. (2022). Modeling reverse auction-based subsidies and stormwater fee policies for low impact development (LID) adoption: a system dynamics analysis. Sustain. Cities Soc. 79:103602. doi: 10.1016/j.scs.2021.103602

Crossref Full Text | Google Scholar

Xia, W., Kong, D., and Zheng, X. (2023). Comprehensive evaluation and spatial layout of high-quality logistics industry development based on accelerated genetic algorithm-projection pursuit model. J. Comput. Methods Sci. Eng. 23, 867–885. doi: 10.3233/JCM-226460

Crossref Full Text | Google Scholar

Yeboah, S.A. (2024). “Bridging the Gap: Public-Private Partnerships in Sustainable Building for Developing Countries”, MPRA Paper, University Library of Munich, Germany. doi: 10.13140/RG.2.2.18576.83206

Crossref Full Text | Google Scholar

Zainuddin, F. A., and Samad, F. A. (2020). A review of crossover methods and problem representation of genetic algorithm in recent engineering applications. Int. J. Adv. Sci. Technol. 29, 759–769.

Google Scholar

Zhang, Z., Chen, Z., Zhang, J., Liu, Y., Chen, L., Yang, M., et al. (2024). Municipal solid waste management challenges in developing regions: a comprehensive review and future perspectives for Asia and Africa. Sci. Total Environ. 930:172794. doi: 10.1016/j.scitotenv.2024.172794,

PubMed Abstract | Crossref Full Text | Google Scholar

Zhao, J., Liu, Y., Zhang, J., Zhang, J., Huang, Y., Yu, L., et al. (2024). An exact method for vehicle routing problem with backhaul discounts in urban express delivery network. Clean. Logist. Supply Chain 11:100157. doi: 10.1016/j.clscn.2024.100157

Crossref Full Text | Google Scholar