Battle royale optimizer with ring neighborhood topology

Abstract

Abstract Recently, battle royale optimizer (BRO), a game-based metaheuristic search algorithm, has been proposed for use in continuous optimization, inspired by a genre of digital games known as “battle royale.” In BRO, each individual chooses the nearest opponent as a competitor. For this purpose, the Euclidean distance between individuals is calculated. This interaction corresponds to an increase in computational complexity by a factor of \(n\). For the purpose of improving the computational complexity of BRO, a modified methodology is proposed using a ring topology, namely, BRO-RT. In the modified version, a set of individuals is arranged in a ring such that each has a neighborhood comprising a number of individuals to its left and right. Instead of a pairwise comparison with all individuals in the population, the best individual among the left and right neighborhoods is selected as the competitor. The proposed scheme has been compared with the original BRO and six popular optimization algorithms. All algorithms are evaluated by applying them to thirteen unimodal and multimodal benchmark optimization functions from CEC2008 and CEC2010. Experimental results show that the BRO-RT algorithm is competitive with or superior to the other seven methods. In addition, the compression spring design problem was used to estimate the ability of the proposed method to solve real-world engineering problems. These results demonstrate that BRO-RT yields promising results when applied to real-world engineering problems. Finally, while BRO is ranked first, and BRO-RT second, they achieved competitive results; BRO-RT has the advantages of lower computational complexity and faster run times than the original BRO algorithm.