A Comparative Study of Two Evolutionary Computation Approaches for Formation of Production Processes in Reconfigurable Manufacturing Systems

Abstract

Due to the capabilities to effectively deal with changing demands and business environments, reconfigurable manufacturing systems (RMS) emerge as a paradigm for manufacturers to respond to challenges in the real world. This is to effectively form a production process based on resources available in RMS to meet the demand, which is an important issue. In this article, we address the process formation problem and model it based on a multi-agent system architecture. To specify the capabilities of machines and robots in RMS, we adopt Petri nets as the modelling tool and formulate the process formation problem as an integer programming problem. Due to the computational complexity in solving the integer programming problem, particle swarm optimization (PSO) and differential evolution (DE) approaches have been applied to solve the problem. Effectiveness of applying these two evolutionary computation algorithms to solve the process formation problem is compared based on experimental results.