A comparative study utilizing hybridized ant colony optimization algorithms for solving dynamic capacity of vehicle routing problems in waste collection system

Abstract

The waste collection stage generated 70% of the cost of the total Municipal Solid Waste (MSW) management system. Therefore, choosing the most affordable waste collection method is essential to accurately estimate the waste collection and transportation costs, thus selecting the required vehicle capacity. The study aims to design a waste collection system for calculating waste collection and transportation costs using a systematic framework that includes Hybridized Ant Colony Optimization (HACO) with Sequential Variable Neighborhood Search Change Step (SVNSCS) and Sequential Variable Neighborhood Decent (SVND). The framework addresses a Dynamic Capacity of Vehicle Routing Problem (DCVRP) and improves ACO's ability in exploration and exploitation stages. The objectives are to minimize the cost of travel distance and arrival time formulated in a mathematical model and to design a new strategy for eliminating the sub-tour problem in the following steps: (1) minimize the number of routes assigned, (2) increase the amount of waste in the vehicle capacity, and (3) define the best amount of waste allowed in vehicle capacity. The waste collection system compared HACO with ACO across four benchmark datasets. The results indicate HACO outperformance ACO at 100%, 91%, 100%, and 87%, respectively. The visualization results demonstrated that HACO has fast convergence and can be considered another essential tool for route optimization in the waste collection system.