การจัดตารางรถบรรทุกขาเข้าและขาออกสำหรับท่ากระจาย สินค้าด้วยวิธีวิวัฒนาการโดยใช้ผลต่าง

Abstract

The efficient distribution center operation is to implement a logistics strategy called cross docking to immediately loaded products from inbound trucks to outbound trucks. The operation involves truck scheduling which includes sequencing and assignment inbound trucks and outbound trucks. The important truck scheduling problem is the assignment of inbound trucks and outbound trucks to the inbound doors and outbound doors and the sequence of all inbound trucks and outbound trucks. This paper proposes method to solve inbound and outbound trucks scheduling in a multi-door cross docking system problem with an exact method to develop a mathematical model by using LINGO 14.0 to process the optimal answers.The objective of a mathematical model is to minimize total completion time or makespan. The completion time (makespan) is considered from the arrival time of the first inbound truck enter to inbound doors for loaded products to the time of last outbound truck returns to the distribution center after completely deliver products to customers which is necessary that the outbound trucks exit the outbound doors as quickly as possible to return to the distribution center within the due date and metaheuristics approach by using Differential Evoluation algorithm (DE) to solve the generated instances for compare the quality results and computing time to find the answers. The experimental results uses 30 generated instances show that LINGO 14.0 can find the optimal solutions within fast computing time in small size instances less than 1 second. Medium size instances took longer computing time use 42.50 minutes to find the answers and LINGO 14.0 could not find optimal solutions in large-scale problems within acceptable computing time. This paper has developed a mathematical model with an exact method by using LINGO 14.0 to process the answers. The experimental results show that LINGO 14.0 can find the optimal solutions in small size problems and LINGO 14.0 could not find optimal solutions in medium and large-scale problems within acceptable computing time of 6 hours. On the other hand, the differential evolution can solve small, medium and large problems by optimizing the parameters based on the FE, F and CR which are 6 settings that the most suitable. The experiment found setting 1 has the most suitable for small problems. While medium problems, setting 4 is the most suitable and large problems setting 4 is the most suitable. The results of the obtained answers were then compared with the exact method. It was found that the DE algorithm was able to find the answer within a reasonable period of time using only the maximum amount of time. 25 seconds for a large problem with Microsoft Visual Studio 2019 and the answer was equal to LINGO 14.0 with % Gap equal to 0. In addition, this study applied the DE algorithm to a real case study problem for the scheduling problem of inbound and outbound trucks for multi-door distribution terminals, which were considered by mathematical models in this study. There is only 1 delivery cycle, therefore, using real case study problem data that there is only 1 delivery cycle per day. The data used to compare the difference are 2 problems: medium and large problems. All the data were then processed for solutions with Microsoft Visual Studio 2019. It was found that the DE algorithm medium problem got 2.74% less answers, as well as the large problem, the DE algorithm got the answer at 2.74%. less than 7.56%