Hybrid Algorithm Optimizes Workflow Scheduling in Fog Computing Environments

A newly proposed hybrid scheduling model combining simulated annealing, improved animal migration optimization, and genetic algorithm operators consistently outperforms standalone evolutionary and migration algorithms in fog-cloud computing environments, according to a study published in Cluster Computing. When evaluated experimentally on the Montage-25 scientific workflow, the proposed model delivered performance improvements over standard genetic algorithms and animal migration optimization that were confirmed to be statistically significant through the Wilcoxon signed-rank test with a p-value below 0.05.

The research addresses foundational infrastructure challenges that have emerged alongside the rapid proliferation of the Internet of Things (IoT). Interconnected smart devices increasingly generate vast volumes of data, which have conventionally been transmitted to centralized cloud servers for processing and storage. However, transmitting and storing this enormous quantity of data incurs significant communication latency and operational drawbacks.

While fog computing has become integral to supporting diverse Internet of Things applications, the growing demand for service quality has introduced multifaceted performance trade-offs. System designers must simultaneously navigate competing constraints involving execution time, communication cost, data latency, operational reliability, and security. Within this distributed computing environment, workflow scheduling has emerged as a critical technical area requiring rigorous exploration, as numerous IoT-based applications depend directly on efficient task mapping. Nevertheless, existing scheduling algorithms deployed in fog environments exhibit clear limitations, operating predominantly as single-objective mechanisms or addressing multi-objective trade-offs across only a narrow range of parameters.

To bridge these identified gaps, the authors developed a multi-objective approach termed Simulated Annealing-Based Improved Animal Migration Optimization (SA-based improved AMO). The architectural novelty of the framework centers on a sequential hybridization strategy that methodically integrates evolutionary search principles with probabilistic optimization techniques. Within this sequential pipeline, genetic algorithm operators are deployed to drive solution diversification across the problem landscape. The framework then employs animal migration optimization to execute migration-based exploration, before applying simulated annealing to achieve fine-grained convergence refinement.

The hybrid algorithm was formulated to optimize three specific performance parameters simultaneously: execution cost, makespan, and virtual machine (VM) utilization rate. In comparative benchmark testing using the Montage-25 workflow, the SA-based improved AMO technique demonstrated superior optimization capabilities relative to both standard genetic algorithms and animal migration optimization across all evaluated metrics.

By systematically combining the exploratory strengths of animal migration optimization, the diversifying capabilities of genetic operators, and the probabilistic refinement of simulated annealing, the sequential strategy achieved reduced computational cost alongside enhanced workflow scheduling efficiency.

Subject of Research: Workflow scheduling optimization in fog computing environments

Article Title: A multi-objective workflow scheduling technique for fog environment inspired by simulated annealing based improved animal migration optimization

Article References: Rathi, S., Nagpal, R., Srivastava, G., & Mehrotra, D. (2026). A multi-objective workflow scheduling technique for fog environment inspired by simulated annealing based improved animal migration optimization. Cluster Computing, 29(16), Article 928. https://doi.org/10.1007/s10586-026-06602-y

Image Credits: AI Generated

DOI: 10.1007/s10586-026-06602-y

Keywords: fog computing, workflow scheduling, simulated annealing, animal migration optimization, genetic algorithms, Internet of Things, Montage-25 workflow, virtual machine utilization, multi-objective, workflow, scheduling, technique