Research on optimization of waste sorting and transportation network in smart cities based on garbage volume prediction

Abstract

As urban populations and economies grow, the amount of municipal solid waste is expected to increase. To address this, waste sorting and collection have become crucial for implementing effective waste classification policies and promoting smart city development. Given the uncertainties in waste volume and the environmental impact of wet waste transportation, this paper presents optimization models for both dry and wet garbage vehicle routing. The goal is to minimize total costs, carbon footprint, and secondary pollution. A random chance constraint is introduced based on decision-makers' preferences, and stochastic theory is used to convert this constraint into a probability density equivalent, improving solution efficiency. To tackle the complexity of waste collection and transportation, an improved genetic algorithm (GA-LS) combining genetic algorithms and a two-layer local search method is proposed. This enhances the exploration ability of the solution space. The model is validated through actual case studies and sensitivity analysis, demonstrating that it can generate near-optimal solutions within an acceptable time frame. The optimization reduces costs, minimizes environmental pollution, and improves residents' satisfaction. Furthermore, it provides a theoretical basis for determining the final wet waste treatment site and optimal transportation routes. Future research will integrate generative deep learning models to improve waste volume prediction accuracy, supporting more efficient and cost-effective smart city decisions.