A proposed enhanced dynamic grey wolf optimization algorithm for cloud load balancing

Key findings from this study

• The study found that EDGWO achieved superior performance on 60% of CEC2020 benchmark functions and placed second on the remaining 40%.
• The researchers demonstrate that EDGWO delivered 25% average improvement in resource utilization and load balancing compared to GWO, PSO, GA, and ABC algorithms.
• The authors report that dynamic elite archive blending and stochastic Levy flight perturbation effectively mitigate convergence stagnation in cloud load balancing optimization.

Overview

The study proposes an enhanced dynamic grey wolf optimization (EDGWO) algorithm designed to improve cloud resource allocation and load balancing under dynamic workloads. The algorithm integrates dynamic elite archive blending to enhance solution diversity and convergence, combined with stochastic Levy flight perturbation to mitigate local optima stagnation.

Methods and approach

The EDGWO algorithm employs dynamic elite archive blending mechanisms and stochastic Levy flight perturbation strategies. Evaluation occurred on the CEC2020 benchmark suite (F1–F10) using fitness metrics including best fitness, mean fitness, and standard deviation. CloudSim simulation compared EDGWO performance against grey wolf optimization (GWO), particle swarm optimization (PSO), genetic algorithm (GA), and artificial bee colony (ABC) algorithms, measuring virtual machine load balance and resource utilization.

Results

EDGWO achieved superior performance on six of ten CEC2020 benchmark functions and ranked second on the remaining four across all evaluated fitness metrics. CloudSim simulation results demonstrated that EDGWO outperformed GWO, PSO, GA, and ABC algorithms in both VM load balance and utilization measures. The algorithm achieved an average 25% improvement in resource utilization and load balancing performance compared to baseline optimization methods.

Implications

The integration of dynamic elite archive blending and Levy flight perturbation addresses fundamental limitations in grey wolf optimization by balancing exploration and exploitation. Enhanced convergence characteristics and reduced stagnation behavior suggest applicability to resource-constrained cloud environments experiencing variable demand patterns. The 25% efficiency gain over established metaheuristic approaches indicates potential value for production cloud infrastructure optimization.

Scope and limitations

This summary is based on the study abstract and available metadata. It does not include a full analysis of the complete paper, supplementary materials, or underlying datasets unless explicitly stated. Findings should be interpreted in the context of the original publication.