Applying a Novel iDEaSm with Maintaining Diversity to Wind Turbine Design Optimization Problem

Wind Turbine Design optimization Problem (WTDOP) is a real-world optimization problem. The WTDOP is also the problem of the Evolutionary Computation Competition 2019 organized by the Japanese Society of Evolutionary Computation. The WTDOP as a single-objective optimization problem has one known characteristic and two reported characteristics. The first characteristic is that the constraint satisfaction of a solution candidate can be obtained only when the solution candidate is evaluated. The second characteristic is that constraint violations occur at a high rate. The third characteristic is that the WTDOP has a multimodal fitness landscape. We propose three improvements to the existing method (new Evaluation Function, Distance Limit, and Deleted Individuals). We compared the results of the numerical experiment based on the median of 21 experiments according to the rules of the competition. The fitness function value of the best solution, the mean, and the median were all better than the existing method. Moreover, in the best result of all trials, diversity was maintained until the final generation.