10-06-2017, 07:14 AM
ABSTRACT
Due to optimal use of processors as well as spending less time, the task scheduling in multiprocessor systems is of great importance. This is one of the NP_hard problems and achieving the optimal schedule or finding the minimum schedule length, using the dynamic algorithm and back-tracking programming, would be time-consuming. Therefore, heuristic methods like genetic algorithms are suitable methods to schedule tasks in a multiprocessor system. In this paper, a new genetic algorithm is presented whose priority of tasks execution is based on the number of their children. The results show that our developed algorithm finds the near-optimal schedule in a reasonable computation time, compared to other heuristics.