Hybridization of the very optimistic method of minimization (VOMMI) algorithm with Taguchi’s orthogonal array for global and engineering optimization

Abstract Motivated by the increasing demand for more efficient optimization techniques, in the present article we propose a simple and efficient hybridized algorithm, the hybrid VOMMI (very optimistic method of minimization), for global and engineering optimization. The standard VOMMI is developed for global optimization based on the social behavior of individuals in deriving inspiration from the successful individuals. However, because of poor exploitation capability, the standard VOMMI suffers from slow convergence to a global optimum. The proposed hybrid VOMMI algorithm is an improved hybrid version of the standard VOMMI algorithm to achieve better computational efficiency. Here, Taguchi’s orthogonal array is used to generate an initial population of standard VOMMI. Then, an orthogonal learning strategy is combined with the optimization mechanism of the standard VOMMI to update the population with further iterative optimization. Numerical experiments are conducted to evaluate the performance of the hybrid VOMMI and to compare it with that of other published algorithms. In the experimentation, a large set of 36 benchmark test functions with varied characteristics is used to examine the validity of the hybrid VOMMI algorithm. Also, five real-world optimization problems are used to test the applicability of the hybrid VOMMI algorithm to engineering design optimization. The statistical analysis of the results obtained from the numerical experiments shows outperformance of the hybrid VOMMI algorithm relative to the other algorithms in terms of its effectiveness, robustness, and computational efficiency for global optimization.