Multi-objective nondominated sorting invasive weed optimization algorithm for the permanent magnet brushless direct current motor design

Abstract

In this paper, we proposed a new multi-objective optimization algorithm named Nondominated Sorting Invasive Weed Optimization (NSIWO) which was inspired from Nondominated Sorting Genetic Algorithm II(NSGAII) and Invasive Weed Optimization (IWO). Firstly, the fast nondominated sorting algorithm was used to rank the weeds, and the number of seeds produced by a weed increased linearly from highest rank to the lowest rank. Moreover, in order to get a good distribution and spread of Pareto-front, crowding distance was used for determining the seeds numbers produced by the weeds with the same rank. Finally, the maximum number of plant population of IWO was adjusted dynamically according to the number of nondominated solutions obtained during each iteration. Then the NSIWO approach was applied to the design of a Permanent Magnet Brushless Direct Current (PMBLDC) Motor of Underwater Unmanned Vehicle (UUV). The obtained results were compared with NSGA-II which is widely used in motor optimization. Numerical results in terms of convergence and spacing performance metrics indicates that the proposed multi-objective IWO scheme is capable of producing good solutions.