Study of the ILS-RVND Algorithm on the Split Load Vehicle Routing Problem with Simultaneous Delivery and Pickup and Its Implementation
Abstract
Distribution companies now face challenges in allocating resources efficiently to maintain competitiveness in an increasingly competitive market. The problem of distributing goods from a depot to customers, where customer demand may exceed vehicle capacity and delivery and pickup are performed simultaneously, can be solved using a variant of VRP called Split Load Vehicle Routing Problem with Simultaneous Delivery and Pickup (SLVRPSDP). State of the art in this article, the SLVRPSDP problem is solved using the Iterated Local Search Random Variable Neighborhood Descent (ILS-RVND) algorithm. There are three stages in the ILS RVND algorithm, namely initialization of the initial solution with Sequential Insertion, improvement of the solution with the RVND algorithm, and acceptance of optimal conditions with a perturbation procedure. The Borland Delphi 7.0 program was designed by creating a flowchart, coding, and implementation with inputs in the form of points as customers, distances between customers, demands and returns, vehicle capacity, ILS iterations, and ILS-RVND iterations. The program’s output consists of the routes formed, total distance traveled, number of vehicles used, and graph visualization. Testing the real-world case study data of goods distribution at China Express compared to the ILS-RVND algorithm on SLVRPSDP produced a solution with a small gap.