Optimized Task Scheduling Using Differential Evolutionary Algorithm

Task scheduling plays a key role for efficiently assigning resources to tasks and performing multitasking. In heterogeneous environments, hard computing task scheduling does not give optimal solution. There are many soft computing techniques used for task scheduling such as evolutionary algorithm which includes genetic algorithm, Differential Evolution (DE), metaheuristic, and swarm intelligence like particle swarm intelligence and ant colony optimization. Genetic Algorithms give locally optimum solution but get stuck in nonoptimal conditions and suffers from quick convergence. DE does not get stuck in local minima and gives a globally optimum solution. Rate of convergence of DE is also slower than GAs and increases with problem size. We have implemented DE for solving task scheduling problem and results demonstrated significant improvement in the fitness of solution with varying parameters as mutation factor, crossover probability, number of iterations, and population. The main aim of this paper is to visualize the effect of variation in various parameters of DE algorithm on the solution of task allocation problem.