Hybrid evolution strategy approach for robust permutation flowshop scheduling

Abstract

In this paper, a robust schedule has been proposed to deal with uncertainities for m-machines permutation flow shop problems. A robust schedule ensures that the expected finish time is always less than the makespan. To use the global search ability of the evolution strategy (ES) and local search ability of Tabu Search (TS), a hybrid evolution strategy (HES) is proposed by combining Improved ES with TS to generate the robust schedules. The robust schedule is first generated using ES and then the solution is optimized using TS for maximum exploitation and exploration of the solution space. For maximum exploitation in ES, (1+9) reproduction operator and double swap mutation is used. Also variable mutation rate is used for fine tuning of the results. In TS, the length of Tabu list is fixed, also lower bound is used to save computational time. The hybrid algorithm is tested on Carlier and Reeves benchmark problems taken from the OR-library. Achieved results are compared with other famous techniques available in the literature, and the results show that HES performs better than other techniques and provides an affirmative percentage increase in the probability that the expected finish time is less than the makespan.