Multi-objective location-routing optimization based on machine learning for green municipal waste management.

Area of Science:

• Operations Research
• Environmental Science
• Computer Science

Background:

• Traditional municipal waste management (MWM) focuses on location-routing problems (LRP) for disposal centers and collection paths.
• Existing MWM systems often overlook balancing economic costs, environmental impact, and resident satisfaction.
• Waste demand variability, modeled as a normal distribution, adds complexity to traditional LRP.

Purpose of the Study:

• To develop a green MWM system optimizing total cost, carbon emissions, and residential satisfaction.
• To address the complexities of multi-objective and non-deterministic factors in waste management.
• To propose an efficient algorithm for solving the green MWM location-routing problem.

Main Methods:

• Modeling the green MWM system as a three-objective location-routing problem.
• Incorporating waste demand as independent discrete random variables following a normal distribution.
• Developing a multi-objective optimization algorithm utilizing a decision tree classifier to guide the search process.

Main Results:

• The proposed algorithm demonstrates high competitiveness against state-of-the-art methods.
• Experimental results validate the algorithm's effectiveness in solving complex MWM problems.
• A case study in Beijing showcases efficient location-routing strategies balancing cost, emissions, and residential satisfaction.

Conclusions:

• The developed algorithm effectively balances total cost, carbon emissions, and residential satisfaction in MWM.
• The decision tree-guided optimization approach enhances efficiency and avoids blind searching in LRP.
• This research provides a robust framework for sustainable and resident-centric municipal waste management.