Authors: G. Subashini; M.C. Bhuvaneswari
Addresses: Department of Information Technology, PSG College of Technology, Coimbatore 641 004, India. ' Department of Information Technology, PSG College of Technology, Coimbatore 641 004, India
Abstract: Both parallel and distributed systems play a vital role in the improvement of high performance computing. A primary issue concerned with the performance of a parallel application executing on a distributed system is allocating the tasks of the application among the various processors in the system. As several conflicting factors influence the allocation strategy, it is necessary to account for multiple objectives. To handle the multi-objective task allocation problem, a Multi-objective Adaptive Particle Swarm Optimisation (MO-ANPSO) with non-dominated sorting is proposed in this paper. The algorithm is implemented and tested on a data set comprising several instances of task interaction graph that models the application. The results show that the proposed method obtains a set of optimal allocations with increased level of performance over the other PSO methods.
Keywords: task allocation; multi-objective optimisation; non-dominated sorting; dynamic inertia; particle swarm optimisation; adaptive PSO; distributed computing; parallel computing; modelling; task interaction graphs.
International Journal of Computer Applications in Technology, 2012 Vol.44 No.4, pp.293 - 302
Received: 08 May 2021
Accepted: 12 May 2021
Published online: 27 Oct 2012 *