การประยุกต์ใช้การหาค่าเหมาะสมที่สุดแบบกลุ่มอนุภาคแบบหลายวัตถุประสงค์สำหรับปัญหาการกำหนดประตูและจัดลำดับท่าเปลี่ยนถ่ายสินค้าแบบหลายประตูรับส่งสินค้า

Abstract

This research aims to solve door assigning and sequencing in a multiple inbound and outbound door cross docking problem to minimize the total earliness and the total tardiness according to Just-In-Time concept. In this study, a mathematical model is developed and solved with LINGO solver in order to find the optimal solution for each objective. The results show that LINGO can find the optimal solutions in small to medium size problem. However, with some limitations, LINGO is not able to attain solutions with multiple objectives simultaneously. Moreover, LINGO could not find solutions in large-scale problem within acceptable computing time. Consequently, this research proposes a metaheuristics called a particle swarm optimization (PSO) in order to deal with large-scale problems. Four encoding and decoding strategies are developed for assigning and scheduling inbound trucks and outbound trucks in order to find quality non-dominate solutions. The numerical experiments of 30 generated instances,varied by problem sizes: small to very large-scale problems, illustrate that PSO show its efficiency in finding Pareto solutions (non-dominate solutions). PSO is able to find optimal or near optimal solution compared with solutions obtained by LINGO. Moreover, PSO can find several non-dominated solutions in a single run within reasonable time. This will be beneficial to final users or decision makers to select a solution based on their preference. The linkage of distribution networks with cross docking and truck sequencing is able to increase the performance and reliability of network due to the uncertainty of lead time of original distribution network.