A Wind Driven Optimization Algorithm for Global Optimization of Electromagnetic Devices

Improvements to overcome premature convergence of existing wind driven optimization algorithms are proposed. The specific measures being proposed include: 1) selection of the origin point of every parcel using dynamic and random tournament selection mechanisms to guarantee that there is a good balance between exploration and exploitation searches. In this paper, the 'so far searched best solution' will be exploited to guide the movement of the randomly initialized parcels by introducing a newly designed mechanism; and a probabilistic mutation is designed; and 2) full utilization of the latest information accumulated from the searched history in order to guide the search toward the potential solutions to enhance convergences, and the 'so far searched worst parcel' is used to shift the new parcel away from the parcel in issue. Numerical results on three case studies are reported to showcase the feasibility and the merit of the proposed method in solving both practical engineering design problems and mathematical test functions.