Basketball team optimization algorithm (BTOA): a novel sport-inspired meta-heuristic optimizer for engineering applications

Abstract

Abstract Real-world optimisation problems are increasingly high-dimensional, nonlinear and constrained. The No Free Lunch theorem implies that no single optimiser dominates across all problem classes, making domain-specific metaheuristics indispensable. Yet mainstream population-based methods often converge prematurely and fail to balance exploration and exploitation under such complexity. To address these limitations, we propose the Basketball Team Optimisation Algorithm (BTOA), a sports-inspired metaheuristic. BTOA maps four basketball concepts-high-intensity training, fast breaks, dynamic positioning and coordinated passing-onto cooperative search operators. In addition, we introduce two extensible modules: (i)a dynamic positioning strategy guided by diagonal structures, significantly improving global exploration capabilities, and (ii) a VariableAttributes to manage the distribution of individual diversity. These modules can be embedded into other population-based optimisers, enriching the heuristic design space. Extensive experiments on the CEC2005 and CEC2017 benchmark suites with 30, 50 and 100 dimensions show that BTOA attains the lowest mean error on 82.61% of the CEC2005 functions and on 66.67%, 63.3% and 66.67% of the CEC2017 functions, respectively. Wilcoxon signed-rank and Friedman tests confirm the statistical significance of these gains. Additional comparisons against several recently proposed algorithms and competition-winning algorithms further highlight BTOA’s consistent advantage. Beyond benchmark tests, BTOA performs well on real-world problems with complex constraints and large decision spaces, such as UAV path planning. Its principled design alleviates key shortcomings of existing metaheuristics and offers a scalable, reliable tool for contemporary engineering optimisation tasks.