A RESTRICTED SIMULATED ANNEALING FOR HYBRID FLOWSHOP SCHEDULING WITH MULTIPROCESSOR TASKS

This paper considers the hybrid flow-shop scheduling problem with multiprocessor tasks which has proven to be NP-hard problem. Due to complexity of the problem, we addressed a restricted simulated annealing (RSA) in which a deduced predecessor-successor relationship of jobs is constructed as a neighborhood search engine. The information of the deduced predecessor-successor relationship of jobs was dynamically updated in the process of SA. The RSA algorithm is experimented on the benchmark instances and compared with existing algorithms in the literature. The computational results show that the RSA algorithm yields very high quality solutions on the same benchmark instances, especially for large problems.