Decreasing the System Testing Makespan in a Computer Manufacturing Company

Abstract

This paper investigates the system testing scheduling problem (STSP), using one of the largest computer manufacturing companies in the world as a case study. A mixed integer linear programming (MILP) model and a restricted simulated annealing (RSA) heuristic which applies two rules to eliminate ineffective job moves to minimize makespan in the STSP are presented. The proposed RSA is empirically evaluated using 188 simulation instances derived from the characteristics of a real technology company. The RSA computational results are compared with those of the traditional simulated annealing (SA) and the artificial bee colony (ABC) algorithms. The statistical results demonstrate that the RSA, SA, and ABC provide much better solutions than the MILP model solved using Gurobi solver for small problems within a reasonable execution time. The RSA offers significant improvements over the SA and ABC algorithms when applied to large problems. The simulation results demonstrate that the proposed RSA heuristic significantly decreases system testing makespan in a computer manufacturing plant at Taiwan.