Date of Publication :5th January 2025

Abstract: Particle Swarm Optimization (PSO) is a popular example of a swarm intelligence technique. PSO has been applied to a variety of fields, including noise, dynamic settings, multi-objective, limited, mini-max, bioinformatics, cloud computing, and medical informatics to mention but a few. Current studies on the PSO algorithm are examined in this paper. Current high-impact studies that have studied and/or modified PSO algorithms have been the focus of the review. The main advantages of the PSO are its ease of use and small number of fine-tuning parameters. The early convergence and lack of a search space balance between exploration and exploitation searches; however, are the main drawbacks of PSO. In this paper, Mathematical operations known as benchmark functions are employed to assess the performance of the algorithm. These functions are complex and possess a range of characteristics. Benchmark functions are used instead of real-world objective functions because they perform reliably better in algorithm testing. Few benchmark functions, spanning from 1 to 50, are listed in most recent literature. Its most important characteristic is the fundamental classification of benchmark functions into unimodal and multimodal functions, each with unique features. In this study, population sizes of 20, 50, 70, 100, and 120 were used, along with two different categories of bench mark functions. A visual depiction of the findings was given.

Reference :