A genetic scheduling algorithm based on knowledge for multiprocessor system

With the extensive studies of the task scheduling problem, many new methods, especial genetic algorithms, have been introduced into this field. In this paper, we develop a novel genetic algorithm, namely the knowledge-based genetic scheduling (KGS) algorithm with task duplication. KGS is different from the previously proposed genetic algorithms in a number of ways. Unlike the others genetic algorithms, KGS initializes population based on more knowledge to provide itself a better iterative basis. KGS also designs an effective decoding algorithm to get the best schedule scheme for a certain chromosome. In addition, KGS uses the relative precedence constraints other than absolute priorities to determine the schedule order of tasks. Simulation results show that KGS outperforms the previously proposed algorithms in terms of the solution quality and the execution time.