Impacts of Scheduling Decisions Based On PSO Algorithm and Dispatching Rules on FMS Performances

Abstract

This paper investigates the scheduling problem in a typical flexible manufacturing system (FMS) with routing flexibility. The scheduling decisions have been established in terms of how the parts are launched into the system, how these parts are routed through various machines and how parts are sequenced for processing on a machine. The first decision is taken using one of the two dynamic dispatching rules (least work remaining (LWR) and most work remaining (MWR)) or first in first out (FIFO). Concerning the second decision, an algorithm based on particle swarm optimization (PSO) is used for real time alternative routing selection of each part whereas shortest processing time (SPT), longest processing time (LPT) and FIFO are used as sequencing rules for working centres (third decision). The simulation results evaluated in terms of the productivity and the utilisation of the system show that the proposed PSO algorithm performs the best when it is combined with LWR as part launching rule and SPT as machine sequencing rule.