Minimizing Total Flow Time in Permutation Flowshop Scheduling by Two-Phase Approach

Abstract

This work considered minimizing total flow time for general n-jobs and m-machines permutation flowshop scheduling problems, which are well-known to be NP-hard. The proposed approach used the combination of Decision Tree (DT) based heuristic algorithm and Scatter Search (SS) algorithm. Unlike existing heuristics available for flowshop scheduling, the DT based algorithm is used to represent the low level data into high level knowledge ie., in the form of IF-THEN else rules which are understandable by the semi-skilled worker. The SS has wide exploration of search space through diversification. The advantages of both DT and SS are utilized to form a two-phase approach. The seed solution generated from DT based approach is taken as input seed to SS algorithm to find good solution. The proposed approach is compared with most effective heuristics and recent meta-heuristics available in the literature. The proposed approach gives better solution as compared to existing composite heuristics and comparable with existing hybrid meta-heuristics.