Solving many-objective reentrant hybrid flowshop scheduling problem considering uncertainty factors in thin-film transistor liquid crystal display

Abstract

The thin-film transistor liquid crystal display (TFT-LCD) front-end array manufacturing process exhibits reentrant characteristics, with uncertainties in the transportation of glass substrates during shop scheduling, further impacting carbon emissions. This study develops a reentrant hybrid flow shop scheduling model considering carbon emissions under uncertain transportation time, where the uncertain transportation time is specifically defined by a triangular fuzzy number (TFN), and a crossing reentrant job handling mechanism is proposed. According to the characteristics of the problem, the shop scheduling process is optimized. In scheduling optimization, the Pythagorean fuzzy set (PFS) is used to solve the problem of uncertain transportation time, and the MYCIN uncertainty factor method, originating from the MYCIN expert system, is employed to evaluate the scheduling scheme and assist metaheuristic algorithm decision-making. A many-objective decision-making method based on PFS and MYCIN uncertainty factors theory is proposed. The golden section factor and the Levy flight are introduced into optimal foraging algorithm (OFA). An improved OFA based on the PFS and MYCIN uncertainty factors (PMYCIN-OFA) is then designed. Finally, three types of experiments are conducted: test cases testing, factory application case testing, and industrial software Flexsim simulations. The results demonstrate that the PMYCIN-OFA surpasses the performance of five classical multi-objective intelligent optimization algorithms and can provide practical solutions in the actual TFT-LCD front-end array manufacturing workshop.