Solving Task Migration Problems byOptimizing the Crossover andMutation probabilities using tabu search

Abstract

Task migration ( TM ) is the act of transferring a process between two processing units, it is considered as an effective strategy to achieve load balancing and higher source utilization, in another way, it is a call to the operating system during run time.TM has been widely used in the distributed systems domain, but as a drawback, the migration time overhead may be long in point that it will increase network congestion and degrade the system performance, we aim to optimize those problems by using genetic algorithms ( GA). Tabu search is a metaheuristic that guides a local search heuristic to escape from local minima and in the same time, to implement an exploration scheme. The simple tabu search algorithm applies a local search where at each iteration, the best solution among the list of neighborhoods is selected and remarked as a new current solution. A short-term memory is implemented as a tabu list where solution attributes are stored to avoid short term cycling. Crossover is the process in which genes are selected from the parent chromosomes and new offspring is created, where the mutation is applied on one chromosome. The genetic algorithms’ performance is largely influenced by crossover and mutation operators, the probabilities of applying those operators can be optimized by tabu search technique probabilities.