Hybrid of Human Learning Optimization Algorithm and Particle Swarm Optimization Algorithm with Scheduling Strategies for the Flexible Job-Shop Scheduling Problem

Haojie Ding,Xingsheng Gu

DOI: https://doi.org/10.1016/j.neucom.2020.07.004

IF: 6

2020-01-01

Neurocomputing

Abstract:The flexible job-shop scheduling problem (FJSP) is a well-known combinational optimization problem. Studying FJSP is essential for promoting production efficiency and effectiveness. Different kinds of improved particle swarm optimization (PSO) algorithms have produced superior results for FJSP in the last few decades. Meanwhile, the human learning optimization (HLO) algorithm, a simple and adaptive learning algorithm for learning system, has helped improve algorithm performance by imitating human learning behavior in recent research. The study proposes a hybrid HLO-PSO algorithm, which utilizes various combinations of the proposed improved PSO and proposed scheduling strategies to solve FJSP under the algorithm architecture of HLO. With the guidance of HLO, the individual learning ability of every particle is further promoted based on the existed advantage of collective action decision of PSO; and with the help of rule-based scheduling strategies, the search capacity of the proposed improved PSO is also further enhanced. By the detailed exposition and analysis, the proposed HLO-PSO is easily implemented and embedded in other production system software or learning system software. Meanwhile, by using it to solve several groups of FJSP instances, the result comparisons with other related algorithms reveal that HLO-PSO can efficiently solve most of single-objective FJSP. (C) 2020 Elsevier B.V. All rights reserved.