Robust optimization model for capacitated vehicle routing problem in waste transportation under demand uncertainty

Abstract

Urban waste transportation systems often experience inefficiencies due to uncertainty in daily waste generation, leading to vehicle overloads, increased operational costs, and environmental impacts. This study proposes a robust optimization model for the capacitated vehicle routing problem (RO-CVRP) to explicitly address demand uncertainty in municipal waste collection. A budgeted uncertainty parameter gamma is incorporated to control the level of protection against worst-case deviations. Initial routes are generated using the Clarke-Wright savings (CWS) algorithm and subsequently evaluated within a robust optimization framework. Computational experiments are conducted using real data from temporary disposal sites (TPS) in Tanah Enam Ratus Subdistrict Medan City, with a vehicle capacity of 10 m³. The results show that higher gamma values produce more conservative routing solutions, increasing the number of vehicles while reducing the risk of capacity violations. Price of robustness (PoR) analysis highlights the trade-off between transportation cost and reliability, confirming the model’s effectiveness for resilient waste logistics planning.