CGA: Chaotic genetic algorithm for Fuzzy job scheduling in Grid environment

We introduce a Chaotic Genetic Algorithm (CCA) to schedule Grid jobs with uncertainties. We adopt a Fuzzy Set based Execution Time (FSET) model to describe uncertain operation time and flexible deadline of Grid jobs. We incorporate chaos into standard Genetic Algorithm (GA) by logistic function, a simple equation involving chaos. A distinguishing feature of our approach is that the convergence of CGA can be controlled automatically by the three famous characteristics of logistic function: convergent, bifurcating, and chaotic. Following this idea,, we propose a chaotic mutation operator based on the feedback of fitness function that ameliorates GA, in terms of convergent speed and stability. We present an entropy based metrics to evaluate the performance of CGA. Experimental results illustrate the efficiency and stability of the resulting algorithm.