Fault-Tolerant Load Balancing Using Enhanced Metaheuristic Algorithms: A Comparative Study with Hierarchical Dragonfly Optimization

Abstract

Load balancing in distributed systems is critical for ensuring optimal resource utilization and fault tolerance. Traditional algorithms often struggle with premature convergence, stagnation in local optima, and ineffective fault recovery strategies. This paper evaluates five nature-inspired metaheuristic algorithms Squirrel Search Algorithm (SSA), Dragonfly Algorithm (DA), Grasshopper Optimization Algorithm (GOA), Grey Wolf Optimizer (GWO), and Butterfly Optimization Algorithm (BOA) for fault-tolerant load balancing. Based on the comparative results, we propose a Hierarchical Dragonfly Algorithm (HDA), which enhances DA by introducing a multi-level adaptive strategy that dynamically adjusts exploration and exploitation during task scheduling. HDA integrates a hierarchical decision-making process to optimize load balancing while incorporating an adaptive fault recovery mechanism that proactively detects failures and redistributes tasks. Simulations on cloud/edge computing environments demonstrate HDA’s superiority in minimizing response time, energy consumption, and service-level violations while handling node failures. Comparative results highlight HDA’s convergence speed, exploration-exploitation balance, and robustness under dynamic conditions, offering a viable solution for modern distributed systems.