Multi-objective Firefly Algorithm for Energy Optimization in Grid Environments


Current researches are focusing on optimizing energy consumption in Grid computing [1], being the job scheduling a challenging task. These researches reduce the energy consumption by heuristics or greedy algorithms and some of them try to balance this reduction regarding the execution time using weights for evaluating these objectives. In this work, a new approach is studied related to the multi-objective optimization for these two conflictive objectives, considering them with the same importance. The obtained solutions show the suitable resources for each job and their order of execution. This new approach is called MO-FA (Multi-Objective Firefly Algorithm) and it is based on the recent FA (Firefly Algorithm)[2] adding multi-objective properties to the preceding versions. The scheduler is implemented in the well-known grid simulator, GridSim to recreate the performance of grid infrastructures and compare MO-FA with other schedulers like Workload Management System (WMS) from the most used European middleware Lightweight Middleware for Grid Computing (gLite) and also the well-known Deadline Budget Constraint (DBC) from Nimrod-G.