Multiobjective Small-World Optimization for Energy Saving in Grid Environments

Job scheduling is a challenging task in grid environments and reducing execution time is one of the most important ways to tackle this problem. Nowadays, society is also more conscious about energy saving, making this fact a really important issue to pursue. In this paper, a new multiobjective algorithm is presented based on the small-world phenomenon: Multiobjective Small-World Optimization (MOSWO), to optimize both objectives: energy consumption and execution time. MOSWO is compared with another recent and swarm algorithm based on the firefly's behaviour: Multiobjective Firefly Algorithm (MO-FA). Both algorithms are compared with the standard multiobjective algorithm Non-dominated Sorting Genetic Algorithm II to show their efficiency as multiobjective approaches. Moreover, the best algorithm proposed, MOSWO, is compared with MOHEFT (multiobjective version of one of the most used algorithms in workflow-scheduling, HEFT) and also with two real grid schedulers: Workload Management System and Deadline Budget Constraint. The results show the advantages of our proposal.