An improved artificial bee colony for multi-objective distributed unrelated parallel machine scheduling

Distributed scheduling has been frequently investigated with the increasing applications of multi-factory production; however, distributed unrelated parallel machine scheduling problem (DUPMSP) is seldom considered. In this study, multi-objective DUPMSP is considered and an improved artificial bee colony (IABC) is presented to minimise makespan and total tardiness simultaneously. Problem-related knowledge is proved and knowledge-based neighbourhood search is proposed. Employed bees and onlooker bees are decided dynamically and not given fixed numbers in the search process. Different combinations of global search and neighbourhood search are used in employed bee phase and onlooker bee phase. A new way is applied to execute scout phase. Extensive experiments are conducted on the effect of new strategies and performances of IABC. Computational results demonstrate that IABC has reasonable and effective strategies and very competitive performances on solving the considered DUPMSP.