Abstract
This study introduces the Morabaraba Optimization Algorithm (MOA), a new global optimization method derived from the sequential and strategic dynamics of the traditional Morabaraba board game. The algorithm translates the main stages of the game into computational search mechanisms, including team allocation, piece placement, movement, flying, mill construction, and cow shooting. In MOA, candidate solutions are assigned competitive roles by separating the population into two rival groups, allowing each solution to adjust its position according to team leaders, board-line structures, and previously generated mill patterns. A new alignment-driven mill formation mechanism is also developed to model strategic player behavior, enabling the algorithm to create mills and weaken the opposing group. The performance of MOA is evaluated on 50 benchmark functions, including unimodal, multimodal, and fixed-dimensional test problems, and compared with 16 established optimization algorithms. The experimental outcomes indicate that MOA achieves rapid convergence while maintaining strong exploration during the early search stages. This behavior is mainly attributed to the integration of mill formation, cow shooting, phase-based position updating, and the structured division of the population into two competing teams. Non-parametric statistical analysis further confirms that MOA provides statistically significant improvements over several competing methods. The results also show that the proposed algorithm performs reliably across a broad set of benchmark functions, demonstrating its robustness and adaptability. In addition, MOA maintains an effective balance between exploration and exploitation, performs consistently in high-dimensional search spaces, and shows strong potential as a Morabaraba-inspired metaheuristic for solving global optimization problems.