Task Allocation in Torus Mesh Networks Using Differential Evolution

Abstract

In the paper, the problem of efficient task allocation in torus mesh network is considered. The authors tested the implemented metaheuristic algorithm which is based on Differential Evaluation method. The focus is taken on tuning the algorithm, i.e., choosing the best parameters of mutation scheme. The research was made using the new designed and implemented experimentation system. Ten mutation schemes were taken into consideration. These schemes and a random algorithm as the reference were compared. The results of experiments showed that using one of mutation schemes called DE/rand/1 can ensure the greatest profit.