A Cooprative Hybrid Evolutionary Algorithm for Flexible Scheduling with AGVs

The flexible job shop scheduling problem (fJSP) is an extension of the traditional job shop scheduling problem (JSP) and is characterized by complexity, stochasticity, and multiple constraints. While evolutionary algorithms (EA) have been used to solve fJSP, the increasing problem scale and the integration of automatic guided vehicles (AGVs) in manufacturing systems present challenges for existing algorithms. This paper proposes a cooperative hybrid EA (ChEA) that uses symbolic and network modelling to represent and solve fJSPs with AGVs. The fJSPs are encoded using a three-stage random-key representation to prevent global optimal deadlocks and ensure solution feasibility. The ChEA approach decomposes the variable and solution spaces into smaller-scale spaces to allow for co-evolutionary optimization. The paper compares the performance of several evolutionary algorithms and identifies particle swarm optimization (PSO) based on Gaussian distribution and locally optimal individuals as the most effective algorithm for global search. The ChEA approach demonstrates competitive performance in terms of average performance, stability, and finding optimal values through numerical experiments.