Hybrid PSO–Aries Metaheuristic Optimization Algorithm for Continuous Optimization

Optimization is critical in the resolution of complex issues within engineering, computer science and artificial intelligence. Most optimization problems found in the physical world can be difficult to solve through direct mathematical methods due to their characteristics which often include non-linearity, high dimensionality and multiple solutions. The use of metaheuristic algorithm techniques have also become very popular amongst the scientific community because of their flexibility and capability to locate near-optimal solutions. This research proposes a hybrid optimization method which combines Particle Swarm Optimization (PSO) with the Aries Metaheuristic Optimization Algorithm. PSO uses a swarm-based methodology that models the social habits of animals flying in flocks (such as birds) and will provide a strong ability to globally explore the solution space. However, PSO is also subject to early convergence and/or slow refinement effects when closing in on the best solution. The Aries algorithm uses an aggressive style that mimics the behavior of warriors during battle and encourages individuals within a community to coalesce towards the best solution while at the same time improving the local searching ability of the individuals in the community. The integration of the global exploratory capabilities associated with the PSO algorithm and the local exploitation ability associated with the Aries algorithm create a stronger overall hybrid search algorithm. The performance of the hybrid PSO-Aries algorithm will be measured using the following benchmark functions: Sphere, Rastrigin, Rosenbrock, Ackley, and Griewank. Experimental tests illustrate that the proposed PSO-Aries hybrid search algorithm will produce improved convergence and better-quality solutions than either the PSO only or the Aries only search algorithms in most cases.