NSGA - II with Controlled Elitism for Scheduling Tasks in Heterogeneous Computing Systems

This paper presents an application of elitist Non-dominated Sorting Genetic Algorithm (NSGA-II), to efficiently schedule a set of independent tasks in a heterogeneous distributed computing system. This scheduling problem is a bi-objective problem considering two objectives. The first objective is minimization of makespan and the second one being the minimization of flowtime. As a multi-objective optimization technique, this paper makes use of Controlled-NSGA-II that modifies NSGA-II to enhance the diversity of solution sets. This algorithm uses a distribution function to control elitism and to get better diversity of individuals. The extent of elitism can be changed by fixing a user-defined parameter. Simulation results on task scheduling problem show the effectiveness of the proposed modification in comparison with NSGA-II.