An effective discrete invasive weed optimization algorithm for multi-AGVs dispatching problem with specific cases in matrix manufacturing workshop. (December 2022)

• Record Type: Journal Article
• Authors: Li, Zhong-Kai; Sang, Hong-Yan; Zhang, Xu-Jin; Zou, Wen-Qiang; Zhang, Biao; Meng, Lei-Lei
• Is Part Of: Computers & industrial engineering. Volume 174(2022)
• Publication Date: 2022-12
• Subjects: Automated guided vehicle -- Dynamic dispatching -- Heuristic -- Discrete invasive weed optimization algorithm; Engineering -- Data processing -- Periodicals; Industrial engineering -- Periodicals; 620.00285
• DOI: 10.1016/j.cie.2022.108755
• Languages: English
• ISSNs: 0360-8352

Abstract

Highlights: A new discrete invasive weed optimization algorithm (DIWO) with rescheduling strategy is proposed. A dynamic rescheduling strategy is proposed. A strategy of setting emergency waiting time is proposed. A heuristic is proposed to dynamically update the task list. An insertion based local search is used to increase the local search ability. Abstract: This paper mainly studies the dynamic multiple automatic guided vehicles dispatching problem (AGVDP) with special cases in matrix manufacturing workshop. The purpose is to introduce dynamic strategy into the static dispatching process to improve the ability to deal with special situations in the scheduling process. To solve this problem, a mathematical model is formulated based on actual production. A new discrete invasive weed optimization algorithm (DIWO) with rescheduling strategy is proposed. A dynamic rescheduling strategy is proposed to improve the ability of scheduling system. A strategy of setting emergency waiting time is proposed based on dynamic rescheduling strategy. A heuristic is proposed to dynamically update the task list, which can effectively update the current task list, solving the problem of task update in special cases and the deletion of the task that has been dispatched. An insertion based local search is used to increase the local search ability. A comprehensive simulation experiment based on the 110 instances from an actual production factory validates the effectiveness of the proposed algorithm.