34 Genetic Algorithm for a Special Class of Multiprocessor Scheduling

The general problem of multiprocessor scheduling can be stated as scheduling a task graph onto a multiprocessor system to minimize the schedule length. In literature, many algorithms have been proposed for the problem, such as the list heuristic, evolutionary algorithm, combine the above two, and so on. But these algorithms not considered part of the tasks needed to be executed on more than one processor simultaneously, so the above algorithms cannot solve or solve not well to this class of scheduling problems. In view of this special class of multiprocessor scheduling, this paper will propose a reasonable genetic algorithm with rational designed encoding and genetic operations. From experiment, we will find the validity of the algorithm proposed in this paper for the multiprocessor scheduling problem in which some tasks need to be executed on more than one processor simultaneously.