Multi-objective constraint task scheduling algorithm for multi-core processors
- 136 Downloads
A task scheduling algorithm is an effective means to ensure multi-core processor system efficiency. This paper defines the task scheduling problem for multi-core processors and proposes a multi-objective constraint task scheduling algorithm based on artificial immune theory (MOCTS-AI). The MOCTS-AI uses vaccine extraction and vaccination to add prior knowledge to the problem and performs vaccine selection and population updating based on the Pareto optimum, thereby accelerating the convergence of the algorithm. In the MOCTS-AI, the crossover and mutation operators and the corresponding use probability for the task scheduling problem are designed to guarantee both the global and local search ability of the algorithm. Additionally, the antibody concentration in the the MOCTS-AI is designed based on the bivariate entropy. By designing the selection probability in consideration of the concentration probability and fitness probability, antibodies with high fitness and low concentration are selected, thereby optimizing the population and ensuring its diversity. A simulation experiment was performed to analyze the convergence of the algorithm and the solution diversity. Compared with other algorithms, the MOCTS-AI effectively optimizes the scheduling length, system energy consumption and system utilization.
KeywordsMulti-core processor Multi-objective constraint Artificial immune Task scheduling
The funding was provided by The Fundamental Research Funds for the Central Universities, Southwest University for Nationalities (Grant No. 2015NZYQN28), National Natural Science Foundation of China (Grant Nos. 11461006 and 11371003) and Special Fund for Scientific and Technological Bases and Talents of Guangxi (Grant No. 2016AD05050).
- 4.Wang, J., Gong, B., Liu, H.: Heterogeneous computing and grid scheduling with hierarchically parallel artificial immune optimization algorithms. ICIC Express Lett. Part B 5, 917–923 (2014)Google Scholar
- 8.Horn, J., Nafpliotis, N., Goldberg, D.E.: A Niched Pareto genetic algorithm for multiobjective optimization. In: Fogarty, T.C. (ed.) Proceedings of the 1st IEEE Congress on Evolutionary Computation, pp. 82–87. IEEE, Piscataway (1994)Google Scholar
- 9.Fonseca, C.M., Fleming, P.J.: Genetic algorithms for multiobjective optimization: formulationdiscussion and generalization. In: International Conference on Genetic Algorithms, pp. 416–423. Morgan Kaufmann Publishers Inc., Burlington (1993)Google Scholar
- 14.Meng, X.F., Xie, W.L.: Research on P2P task scheduling with multi-objective constraints based on immune algorithm. Acta Electron. Sin. 39(1), 101–107 (2011)Google Scholar
- 16.Lee, Y.C., Zomaya, A.Y.: An artificial immune system for heterogeneous multiprocessor scheduling with task duplication. In: International Parallel and Distributed Processing Symposium. DBLP, pp. 1–8 (2013)Google Scholar
- 18.Chou, C.L.: Marculescu. FARM: fault-aware resource management in NoC-based multiprocessor platforms. In: IEEE Design, Automation and Test in Europe Conference and Exhibition, pp. 1–6 (2011)Google Scholar
- 19.Dick, R.P., Rhodes, D.L., Wolf, W.: TGFF: task graphs for free. In: IEEE Proceedings of the Sixth International Workshop on Hardware/Software Codesign, 1998 (CODES/CASHE ’98), pp. 97–101 (1998)Google Scholar
- 20.Duenha, L., Guedes, M., Almeida, H., et al.: MPSoCBench: a toolset for MPSoC system level evaluation. In: IEEE International Conference on Embedded Computer Systems: Architectures, Modeling, and Simulation, pp. 164–171 (2014)Google Scholar