A Load Balancing Using Multi-population Grasshopper Optimization Approach for Workflow Tasks in Clouds

Infrastructure services are offered on demand in a cloud environment using a pay-as-you-go model. Workflow systems are widely used in project management, earthquake analysis, genome processing, large business applications, IT industries, supply chain management, and many other areas. It has already been established that load-balanced workflow process allocation in cloud is NP-hard problem. In this work, a novel method for solving the load balancing problem in the cloud environment is proposed offering an appropriate allocation plan by using meta-heuristic algorithms, namely, grasshopper optimization algorithm (GOA) to maximize the resource utilization. This approach is employing the multi-population (MP) policy for exploring more diversity in the solution space to reduce the load imbalance on virtual machines. The scheduling workflow load balancer is simulated in the MATLAB to accomplish the performance analysis. The results show that the proposed MPGOA outperforms PSO in all the task and VM sets considered in the performance evaluation.