Kabaddi Game Optimizer: A Sport-Inspired Novel Metaheuristic Algorithm

This is the authors' abstract. We don't add key points for this paper.

This paper introduces the Kabaddi Game Optimizer (KGO), a population-based metaheuristic for single-objective minimization inspired by the sport of Kabaddi. The KGO adopts a population-based framework in which individuals are divided into two competing teams. Each team consists of players ranked by fitness and guided by a dynamic leader. During each iteration, players update their positions using two adaptive strategies governed by a probabilistic factor, mimicking raiding and defending behaviors. The algorithm maintains a balance between exploration and exploitation through continuous rank updates, leader exchange, and replacement of weak individuals to avoid premature convergence. The algorithm is developed using MATLAB software and is tested using 40 benchmark functions in addition to CEC 2017, CEC 2020 and CEC 2021 benchmark sets in order to examine its efficiency and reliability. The proposed algorithm is compared against popular optimization methods such as PSO, SaDE, L-SHADE, SSA, WOA, DA, Chimp, and GWO. The comparison is carried out on the basis of several evaluation parameters such as fitness means and median, standard deviation, and convergence time. Results indicate that the KGO algorithm is efficient and can achieve the optimum or better solutions for many of the benchmark functions by taking first place in 28 functions out of 40 functions and proving to be superior to other compared algorithms according to statistical tests. In addition to this, practical feasibility of the suggested algorithm is validated through implementation in four real-world engineering design optimization problems.