Research Of Task Scheduling Based On Heterogeneous Multi-processor System

As one of the highlights in the multi-core processor system, heterogeneous multi-core processor system becomes widely popular and the most commonly used in embedded system, thanks to its high efficiency and low cost characteristics. In the course of the study of heterogeneous multi-core processor system, tasks scheduling is particularly important, because a good algorithm for tasks scheduling can bring system performance into full play.Tasks scheduling for Heterogeneous multi-core processor system have been proved to be a NP complete problem, so to solve such problem some commonly used heuristic algorithms seem inadequate. Then, people turn their attention to intelligent algorithm so as to gain the solution. In this paper, the improved genetic algorithm among intelligent approximation algorithm is applied to tasks scheduling for heterogeneous multi-core processor system. What's more, we get a relatively good solution.This paper first proposes a tasks scheduling mathematical model of heterogeneous multi-core processor system. Besides, on the basis of the basic genetic algorithm, it proposes dynamic genetic algorithm based on good population for tasks scheduling model of heterogeneous multi-core processor system. In order to provide a good basis for computing operation, improved genetic algorithm uses heuristic algorithm to build relatively good population at the time of the population initialization. In the process of algorithm computing, the corresponding fitness function is put forward in connection with tasks scheduling for heterogeneous multi-core processor system. In addition, dynamic adaptive crossover and mutation rate( ) is proposed according to that different population has different individual fitness. As a result, the algorithm becomes more targeted during the process of improving population. In order to try to save the algorithm running time, the population fitness similarity ( ) is added to optimize the algorithm in the settings of algorithm termination condition. Through all these improvements, dynamic genetic algorithm based on the good population not only satisfies the solution requirements, but also has good convergence.For the purpose of verify the performance of improved algorithm, this paper implements this algorithm. Moreover, the corresponding experiments on the feasibility, parameter settings and comparison of similar algorithms are done under the environment of Microsoft Visual C++ 2010. After that, this paper gives the result analysis and finally proves that the improved genetic algorithm can effectively solve tasks scheduling problem for the heterogeneous multi-core processor system.