Ba-PSO: A Balanced PSO to solve multi-objective grid scheduling problem

In a computational grid, environment is dynamic in nature and has distributed resources spread across multiple administrative domains. Therefore, it becomes necessary to provide an effective scheduling mechanism for the applications submitted to the computational grid. Particle Swarm Optimization (PSO) is very popular meta-heuristic in finding solutions to complex problems. Compared to other meta-heuristics, PSO has less parameters and better computational efficiency. In the paper, an advanced form of PSO i.e. balanced PSO (Ba-PSO) has been proposed to solve the scheduling problem of computational grid. The proposed algorithm decreases the jobs' execution time and improves utilization of resources. The proposed, Ba-PSO, is scalable and works for small as well as large datasets. The role of a standard dataset is significant in testing a new algorithm because it helps in investigating the working of algorithm and provides important insights about the algorithm being tested. This paper uses a standard dataset generated by Czech National Grid Infrastructure i.e. Metacentrum. The proposed Ba-PSO algorithm is evaluated using the standard dataset and its results outperforms other considered deterministic and heuristic approaches.

Automated summary

This paper proposes a new balanced PSO algorithm to solve the scheduling problem of computational grid. The algorithm is evaluated using a standard dataset, and its results outperform other considered deterministic and heuristic approaches.