A Review Study of Modified Swarm Intelligence: Particle Swarm Optimization, Firefly, Bat and Gray Wolf Optimizer Algorithms

Abstract

Background: Limitations exist in traditional optimization algorithms. Studies show that bio-inspired alternatives have overcome these drawbacks. Bio-inspired algorithm mimics the characteristics of natural occurrences to solve complex problems. Particle swarm optimization, firefly algorithm, bat algorithms, gray wolf optimizer, among others are examples of bio-inspired algorithms. Researchers make certain assumptions while designing these models which limits their performance in some optimization domains. Efforts to find a solution to deal with these challenges leads to the multiplicity of variants. Objective: This study explores the improvement strategies in four popular swarm intelligence in the literature. Specifically, particle swarm optimization, firefly algorithm, bat algorithm, and gray wolf optimizer. It also tries to identify the exact modification position in the algorithm kernel that yielded the positive outcome. The primary goal is to understand the trends and the relationship in their performance. Methods: The best evidence review methodology approach is employed. Two ancient but valuable and two recent and efficient swarm intelligence, are selected for this study. Results: Particle swarm optimization, firefly algorithm, bat algorithm, and gray wolf optimizer exhibit local optima entrapment in their standard states. The same enhancement strategy produced effective outcome across these four swarm intelligence. The exact approach is chaotic-based optimization. However, the implementation produced the desired result at different stages of these algorithms. Conclusion: Every bio-inspired algorithm comprises two or more updating functions. Researchers need a proper guide on what and how to apply a strategy for an optimum result.