An Iterative Random PSO to minimize Total-flow- time in Flowshop Scheduling

Flowshop scheduling is an interdisciplinary challenge of addressing major optimality criteria of minimizing makespan and total-flow-time. The enumerations for finding the probabilities for improving the utilization of resources turn this problem towards NP-Hard. Particle Swam Optimization (PSO) has proven its ability to find of near optimal solution when problem size is large. Local search techniques increase ability of PSO and prevent PSO from getting stuck into local optima. This paper proposes modification in PSO an Iterative Random PSO without use of any local search technique. It uses re-initialization of population at every iteration; due to which there is no possibility of getting stuck into local optima. The proposed method achieves total-flow-time objective of flowshop scheduling. Computational results are obtained with 110 benchmark instances of Taillard and compared for the totalflow-time criterion. Ultimately, 80 instances out of 110 best known solutions provided by DPSOVNDwere improved by IRPSO. There is still scope of improving IRPSO for dataset of 5 machines, it gives near optimal solution but not better one. Keywords— Metaheuristics; particle swarm optimization (PSO); iterative random PSO (IRPSO); flowshop scheduling; total-flowtime(tft), optimization.