HICSO Algorithm for Performance and Cost Optimization Based Workflow Scheduling in Cloud Computing

Abstract

Cloud workflow scheduling in heterogeneous environments is a challenging optimization problem due to the large task–VM search space and the need to achieve efficient resource utilization. This study proposes HICSO (Hybrid Inertia-weight-based Cat Swarm Optimization), an improved Cat Swarm Optimization algorithm incorporating a proposed Hybrid Inertia Weight (HIW) strategy to improve the balance between exploration and exploitation. HEFT is used to generate the initial population, while HIW is integrated into the tracing mode of CSO. The proposed approach jointly optimizes makespan and cost using a weighted fitness function. Experiments are conducted in CloudSim 3.0 using four scientific workflows—CyberShake, Montage, Inspiral, and Sipht—with 1000 tasks each and 70 heterogeneous VMs. The algorithms are executed for 15 independent runs for comparative evaluation. Results show that HICSO achieves the lowest mean makespan and mean cost across all four workloads, demonstrating improved scheduling efficiency and cost effectiveness compared with LDCSO, CICSO, OICSO, and conventional CSO.