The adaptive wind driven optimization and its application in electromagnetics

In this paper, we analyze in detail the properties of the newly developed Adaptive Wind Driven Optimization (AWDO) method and demonstrate its efficiency on continuous-valued electromagnetic (EM) problems by comparing it to the performance of the classical Wind Driven Optimization (WDO) method. While WDO has proven to be competitive with similar algorithms such as Particle Swarm Optimization (PSO), there are a large number of inherent parameters in the WDO update equations which the user needs to fine tune for optimal performance. The newly developed AWDO method automatically determines these parameters, creating an adaptive algorithm which eliminates the need for the user to have apriori knowledge of the optimal values for the inherent parameters. Numerical optimization results on EM problems indicate that AWDO matches or surpasses the performance of WDO, yielding a self-adaptive algorithm which is both extremely efficient as well as easy to use.