Optimizing Generalized Capacitated Vehicle Routing Problem Using Augmented Savings Algorithm

Abstract

This article presents an Augmented Savings Algorithm (ASA) for solving Generalized Capacitated Vehicle Routing Problem (GCVRP), for single-trip homogeneous and heterogeneous fleet. The ASA is a modified version of classical heuristic, based on savings algorithm for homogeneous fleet, pioneered by Clarke and Wright back in 1964 and further enhanced by others. In the ASA algorithm, we introduced a changed modality for adjusting savings value upon prioritizing the parameters for compactness, distribution asymmetry, and nodal demand. Our approach is verified with respect to the real-time vehicle scheduling of a company bus service in Mumbai. This is to ideally redesign the sub-routes which are embedded in existing routes. The algorithm is further validated with regard to the benchmark instances in the literature. The solutions obtained minimizes the overall cost, i.e., the fixed cost; and the variable cost, after maximizing the occupancy of the pickup vehicles. The ASA approaches besides showing the improvement in the results obtained by others, also demonstrates better results compared to the enumerative parameter setting approach proposed by Altinel and Öncan [2], and Empirically Adjusted Greedy Heuristic (EAGH) approach adopted by Corominas et al. [8].