A Combinatorial Multi-objective Particle Swarm Optimization Based Algorithm for Task Allocation in Distributed Computing Systems
In a distributed computing system (DCS), the scheduling of tasks comprise of two phases, task allocation and task scheduling. The allocation of tasks to different processors is required to maximize the processors synergism in order to achieve the various objectives, such as, system throughput, reliability maximization and cost minimization. The task allocation also need to satisfy a set of system constraints related to memory and link capacity. This problem has been shown as NP-Hard. Most of meta-heuristic algorithms deal with the task allocation problem as single objective or transform the multiple objectives into single objective. This paper presents a combinatorial multi-objective particle swarm optimization based (CMOPSO) algorithm to deal with the multiple conflicting objectives of task allocation problem simultaneously without transforming them to single objective. The performance of the algorithm is compared with a NSGA-II based task allocation algorithm; results manifest that the algorithm shows good performance under different problem scales and task interaction density.
KeywordsTask allocation MOPSO Combinatorial optimization problem Distributed computing system
Unable to display preview. Download preview PDF.
- 2.Coello Coello, C.A., Lechuga, M.S.: MOPSO: a proposal for multiple objective particle swarm optimization. In: Proceedings of the Evolutionary Computation on 2002, CEC 2002, pp. 1051–1056. IEEE Computer Society, Washington, DC, USA (2002)Google Scholar
- 3.Ernst, A., Hiang, H., Krishnamoorthy, M.: Mathematical programming approaches for solving task allocation problems. In: Proceedings of the 16th National Conference of Australian Society of Operations Research, Australia (2001)Google Scholar
- 9.Mostaghim, S., Teich, J.: Strategies for finding good local guides in multi-objective particle swarm optimization MOPSO. In: Proceedings of the IEEE, Symposium on Swarm Intelligence, pp. 26–33 (2003)Google Scholar