An Enhanced Manta Ray Foraging Algorithm with Lévy Flight and Heuristic Operator for Efficient Scientific Workflow Scheduling in Cloud Environments

Abstract

Scientific workflow scheduling in cloud environments is challenging due to the dynamic availability of resources and interdependence among tasks. This article presents a novel hybrid solution—namely Lévy-Heuristic Manta Ray Foraging Optimization Algorithm (LH-MRFOA)—to tackle these challenges. The suggested approach incorporates Lévy flights into the classic Manta Ray Foraging Optimization (MRFO) to enhance global exploration and introduces a heuristic dependency management operator for task assignment optimization and wait time minimization. Comprehensive experimentation on benchmark standards (Inspiral, CyberShake, Montage, SIPHT, and Epigenomics) validates that LH-MRFOA consistently outperforms traditional meta-heuristics (GA, PSO) and variant MRFO enhancements in minimizing makespan and operational cost. These findings recognize LH-MRFOA's potential for large-scale, data-heavy applications requiring timely and affordable use of resources in modern cloud data centers.