An Improved Multi-Objective PSO Algorithm for Workflow Scheduling in Cloud by Improving the

Abstract

An important issue in cloud computing is workflow scheduling. Workflow scheduling is a way to map each work in a workflow into existing processing resources, so that its workflow rules are followed. Proper scheduling can improve the performance of servers and processors in doing tasks. There are several scheduling methods available to optimize the efficiency of servers throughput in the cloud environment. This paper presents an improved evolutionary algorithm for multi-objective scheduling of workflows in the cloud using improving the initial population. In this method, two IC-PCP and LOSS algorithms are used to generate good solutions, and these solutions are placed instead of random solutions in the initial population. Then the initial population obtained is applied to the PSO evolutionary algorithm. The goal of this improved multi-objective PSO algorithm is to minimize the overall cost of executing a workflow, provided that the deadline constraint is fulfilled. The results from our study show that our algorithm generally outperforms the other algorithms.