An adaptive PSO-based real-time workflow scheduling algorithm in cloud systems

Abstract

Cloud computing has emerged as a powerful platform for providing computing resources in the past decade. Developing workflow scheduling algorithms can efficiently reduce the cost of executing tasks in cloud systems. The features of elasticity and heterogeneity of cloud computing bring challenges for scheduling strategies. For real-time workflows, reducing execution time and reducing execution cost are two conflicting objectives. To address this issue, we propose in this paper an improved real-time workflow scheduling algorithm based on particle swarm optimization (PSO). Different from traditional scheduling heuristics which rely on the initial resource pool, our algorithm can adaptively optimize the resource usage. Simulation experiments are conducted to evaluate our algorithm on workflows with different sizes under various deadlines. Compared with the best algorithm ever known, our algorithm shows excellent performance in both cost and makespan.