A Computational Study of Shifting Bottleneck Procedures forShop Scheduling Problems

  • Authors:

  • Ebru Demirkol;Sanjay Mehta;Reha Uzsoy

  • Affiliations:

  • School of Industrial Engineering, Purdue University, West Lafayette, IN 47907;i2 Technologies, Irving, TX;School of Industrial Engineering, Purdue University, West Lafayette, IN 47907

  • Venue:
  • Journal of Heuristics
  • Year:
  • 1997

Quantified Score

Hi-index 0.00

Visualization

Abstract

We examine the performance of Shifting Bottleneck (SB) heuristics forshop scheduling problems where the performance measure to beminimized is makespan (C_{max}) or maximum lateness(L_{max}).Extensive computational experiments are conducted on benchmarkproblems from the literature as well as several thousand randomlygenerated test problems with three different routing structures andup to 1000 operations. Several different versions of SB are examinedto determine the effect on solution quality and time of differentsubproblem solution procedures, reoptimization procedures andbottleneck selection criteria. Results show that the performance ofSB is significantly affected by job routings, and that SB withoptimal subproblem solutions and full reoptimization at eachiteration consistently outperforms dispatching rules, but requireshigh computation times for large problems. High quality subproblemsolutions and reoptimization procedures are essential to obtaininggood solutions. We also show that schedules developed by SB tominimize L_{max} perform well with respect to several otherperformance measures, rendering them more attractive for practicaluse.