A Hybrid Algorithm for Flow Shop Scheduling Problem with Unavailable Time Periods and Additional Resources

Abstract

In the scheduling literature, the studies that consider unavailable periods (UPs) have generally ignored the resources. However, when the resources to be used in unavailable periods are limited and these resources are needed for more than one machine at the same time, the problem of when the resource should be allocated to which machine arises. This decision is important as it can greatly affect the effectiveness of the machine schedule. For this reason, it is necessary to consider not only the UPs, but also the resources used by the UPs. In this study, flow shop scheduling problem with unavailable periods, flexible in a time window, and additional resources is discussed. In the considered problem, since additional resources are required during the unavailable periods and they can serve just one machine at a time, they cannot overlap. A MIP model and a hybrid algorithm that genetic algorithm and modified subgradient algorithm works together, have been developed for the considered problem. The performance of the hybrid algorithm is compared with pure genetic algorithm and Cplex solver of GAMS by using randomly generated test problems. Test results showed that while hybrid algorithm has solution quality advantage, genetic algorithm has solution time advantage. In addition, with the developed hybrid algorithm, GAMS results were improved up to 88%.