A Fast Parallel Dynasearch Algorithm for Some Scheduling Problems

In this paper we propose a parallel dynasearch algorithm to solve the single machine total weighted tardiness problem. Nowadays dynasearch method is the most powerful heuristic method for the single machine total weighted scheduling problem. The neighborhood is generated by executing series of swap moves. Using dynamic programming such a neighborhood, which has exponential size, is explored in polynomial time. We propose parallel dynasearch algorithm for the parallel mainframe computers which significantly improves results of the sequential dynasearch algorithm