Application of a hybrid generation/utility assessment heuristic to a class of scheduling problems

摘要

A two-stage heuristic solution approach for a class of multiobjective, n-job, one-machine scheduling problems is described. Minimization of job-to-job interference for n jobs is sought. The first stage generates alternative schedule sequences by interchanging pairs of schedule elements. The set of alternative sequences can represent nodes of a decision tree; each node is reached via decision to interchange job elements. The second stage selects the parent node for the next generation of alternative sequences through automated paired comparison of objective performance for all current nodes.