Genetic algorithm with reinforcement learning for optimal allocation of resources in task scheduling

Task scheduling in cloud computing is one of primary research problem in computer science technology. Finding an optimal solution for task scheduling not only enhances the system/machine performance but also it reduces the total processing cost. There are number of task scheduling algorithms developed by previous researchers, but none of them have been globally accepted because of their own pros and cons. The current study made an attempt to solve task scheduling for optimal utilisation of virtual machines to execute the assigned tasks with the help of genetic algorithm (GA). The proposed strategy primarily focused on the minimisation of task execution time by considering it as fitness function while GA implementation. Reinforcement learning is integrated with the proposed algorithm in order to enhance its performance while finding the optimal resource allocation. Later, the methodology is validated in several cloud environments with simulated analysis results in order to check its feasibility. Copyright © 2024 Inderscience Enterprises Ltd.