GEAR TRAIN DESIGN SOLUTIONS VIA HYPERSPHERE DYNAMICS DRIVEN ADVANCED PARTICLE SWARM OPTIMIZATION

Abstract

This work presents an advanced version of the Particle Swarm Optimization (PSO) algorithm, a well-known optimization algorithm for the solution of the global optimization problems, called PSO with Hypersphere Dynamics and Mutation (PSO-HDM), to deal with the optimization obstacles. The novel method employs a novel technique where the particles’ positions are updated using the rotation of the hyperspheres, providing for better exploration of the search space. In addition, two new mutation techniques, Jitter and Gaussian, are used to keep away from the local optima and enhance the solution variety. Dynamic modifications of the classical PSO’s parameters, such as cognitive and social coefficients, also improve the algorithm’s achievement. The PSO-HDM optimization algorithm is evaluated with utilizing some benchmark functions and compared to classical PSO, getting better values in determining the optimal solutions. Gear train design problems are selected as an engineering design problem to show the effectiveness of the new suggested method. The obtained results present the capability of the proposed method. This proposed optimization algorithm could be seen as an alternative method to other optimization algorithms proposed in the literature.