A Hybrid Approach Combining Ant Colony Optimization and Simulated Annealing for Cloud Resource Scheduling

Abstract

Cloud computing is imperative to schedule efficiently for tasks and resources to assure performance, reduction of costs, and service-level agreement. Traditional methods cannot balance this complexity, resulting in the conception of a hybrid model that will be based on the integration of ACO with SA. Here, the former algorithm applies the collective intelligence of ACO, coupled with positive feedback, to achieve better quality solutions and escapes the local optimum of the latter. This algorithm runs in two phases: The initial solution is generated using the ACO, and the solution is refined using SA. Simulated experiments within a simulated cloud environment of CloudSim showed that this hybrid approach succeeds in minimizing makespan, reducing energy consumption, and optimizing the cost for different workloads. The ACO-SA algorithm heralds a promising direction toward highly efficient management of cloud resources and opens up further research directions in hybridizing other complementary algorithms.