A hybrid data envelopment analysis_decision tree approach to evaluate the bi-criteria flow shop with blocking problem

Abstract

The purpose of this paper is to evaluate the overall performance of a bi-objective scheduling problem using a method based on a combination of data envelopment analysis and decision tree. The hybrid approach is applied to the blocking flow shop scheduling problem with both makespan and total tardiness objectives. For this end, we propose two families of three-phase heuristics. In phase one, several constructive heuristics are used to generate initial solutions and in phase two, two improving heuristics are used to generate non-dominated solutions. One of the improving heuristics is based on the makespan criteria and the other one is based on the total tardiness criteria. The non-dominated solutions are then compared using the free disposal hull (FDH) formulation of data envelopment analysis (DEA). The overall performance of the composed heuristics is measured using a decision tree approach.