Design of Efficient Combinatorial Workflow Scheduling Algorithm for Cloud Computing

Abstract

Cloud computing, a fast expanding technology, deliveres services on demand via In- newlineternet. In recent times it is gaining popularity in information technology due to the advance- newlinements like on-demand processing, resource sharing, pay per use, etc. There are several cloud newlinecomputing issues like security, QoS management, data center energy consumption, scaling, newlineetc. Scheduling is among the several challenging problems, where several tasks need to newlinebe assigned to resources to optimize the quality of service parameters. This will require a newlinesuitable scheduling algorithm. newlineWorkflow scheduling is considered an NP-hard problem and has a significant issue in newlinethe cloud environment. Finding solutions in the workflow scheduling is even difficult com- newlinepared to task scheduling as the tasks are dependent on each other in the workflow. Finding newlinethe polynomial-time solutions for workflow scheduling problems is difficult with most of newlinethe existing algorithms designed for traditional computing platforms. Some existing meta- newlineheuristics algorithms proposed for workflow scheduling problems are stuck in the local op- newlinetimal solution and fail to give the global optimal solution. newlineIn this thesis, a new single objective model-based hybrid metaheuristic algorithm, newlinenamed HPSOGWO, is proposed. This is the hybrid version of two well-known algorithms, newlineParticle Swarm Optimization (PSO) and Grey Wolf Optimisation (GWO). The proposed al- newlinegorithm takes advantage of the standard PSO and GWO algorithms and does not stick to newlinethe local optimal solution. The proposed algorithm (HPSOGWO) was tested to reduce the newlinetotal execution cost and total execution time of the dependent tasks in the cloud computing newlineenvironment. The experiment results show that the HPSOGWO outperformed compared to newlinethe standard PSO and GWO algorithm in total execution cost and total execution time. newline