ABSTRACT
Biologically inspired optimization techniques are applied in various fields for finding a quick optimal or near optimal solution to computationally hard problems. Study on the application of these techniques for task scheduling in cloud computing systems is worth due to the dynamic and unpredictable nature of task arrival. The quality of a task schedule in cloud computing systems depends on the response time that a user of the cloud system experiences and the cost of the service at both cloud service provider’s end and at the user side. Optimizing such coinciding factors in a dynamic environment is challenging since it is a multi-objective optimization problem. Many of the scientific and large-scale applications are characterized by a huge number of homogeneous tasks, that can be executed in parallel. This work considers the scheduling of such tasks. Three swarm intelligence algorithms namely Particle Swarm Optimization (PSO) technique, Artificial Bee Colony (ABC) algorithm and Ant Colony Optimization (ACO) algorithm and an evolutionary computation-based algorithm namely Genetic Algorithm (GA) are applied to solve the problem. Experiments are done on different types of data sets and a comparative study on the results are performed to determine the suitability of these algorithms in the cloud environment. The results are also compared with a list-based heuristic search algorithm namely Greedy Randomized Adaptive Search Procedure (GRASP).
