High-dimensional function optimization with a self adaptive differential evolution

A good optimization algorithm must be capable of handling high-dimensional problems, meaning that there are many decision variables to be optimized at the same time. The problems of this category are challenging. This paper tests the scalability of wDE, which is a differential evolution algorithm with self-adaptive parameters. The statistical results and convergence graphs from the experimentation using benchmark problems of 100-, 500-, and 2000-dimensions are analyzed and compared to three standard variants of differential evolution algorithm.