Artificial Orca Optimiser: Theory and Applications for Global Optimisation Problems

Abstract

With the growing complexity of real-world engineering optimisation problems, interest in meta-heuristic algorithms is increasing. However, existing meta-heuristic algorithms still suffer from several shortcomings, including a poor balance between global and local search, a tendency to converge toward the centre of the solution space, and susceptibility to getting trapped in local optima. To overcome these shortcomings, a novel meta-heuristic algorithm, called artificial orca optimiser (AOO), is proposed based on the unique behaviours of orcas in nature. Within the framework of AOO, the switching factor, guidance phase, and iterative formulas that do not converge toward the centre of the solution space, are designed to enhance the equilibrium between exploration and exploitation, ensure agents the ability to escape from the local optimum, and comprehensively explore the solution space without being limited to the centre of the solution space, thereby increasing the likelihood of finding the global optimal solution. Qualitative, quantitative, scalability, sensitivity, and practical application analyses of the experimental results demonstrate that AOO overcomes the issue of converging to the centre of the solution space, alleviates the problems of poor balance and susceptibility to the local optimum, and exhibits outstanding optimising performance, fast convergence, great scalability, high robustness, and excellent practicality.