Markov Stochastic Bregman Divergencive Ant Colony Resource Optimized Task Scheduling In Cloud

Abstract

The Cloud Computing (CC) environment enables on-demand network access to a shared resource pool. With the given constraints of resources, task scheduling becomes an essential in CC environment. The computing resources are said to be allocated according to the needs and requirements of users. Besides the exponential increase of users in CC environment, resource allocation has become a tedious process. To address this issue, a novel Markov Stochastic Bregman Divergence based Multi-criteria Ant Colony Resource Optimized Task Scheduling (MSBD-MACROTS) technique is introduced. Initially, the task scheduler receives multiple heterogeneous tasks from the users and applies the Markov Stochastic Bregman Divergence based multi-criteria Ant Colony optimization to find the resource optimal virtual machine for assigning the tasks. The local optimum is selected by sorting the virtual machine based on the fitness measure. Besides, the Markov Stochastic Transition probability obtains better convergence, therefore reducing the overhead involved in task scheduling. Finally, a measure for strength of pheromone update that is based on Bregman Divergence distance is proposed to find global optimum solution in search space. In this way, the entire heterogeneous tasks from users are correctly scheduled to virtual machine. The results are shows that MSBD-MACROTS technique improves the task scheduling efficiency with less computation overhead and memory consumption.