Three-Part Genetic Algorithm to Optimize the Outbound Train Loading Process Using a Multiple Travelling Salesman Problem Approach

Abstract

This work addresses the inefficiencies in the container outbound loading process, focusing specifically on the Wagon Container Assignment Problem (WCAP). Inefficiencies in this area have a significant negative impact on productivity and operational efficiency, leading to increased fuel consumption and higher monetary costs. Given the critical role dry ports play in the global multimodal supply chain, these inefficiencies are particularly concerning. The rising volume of containerized freight underscores the need for optimized cargo handling. Effective management is essential to ensure smooth transitions, enhance terminal performance, and maintain the environmental sustainability of the logistics industry. As a consequence, this work introduces a Three-Part Genetic Algorithm (TPGA) to minimize the total distance travelled to fill an outbound train. Two variations of the algorithm (TPGA1 and TPGA2) were compared with an exact one and two heuristics. The TPGA methodology consistently outperformed heuristics and the matched exact algorithm in the first seven scenarios, demonstrating the best results in distance minimization. TPGA2 demonstrated greater time efficiency for larger scenarios, although TPGA1 performed better with two vehicles.