Publications

Abstract : Differential Evolution (DE), the real parameter optimization algorithm for population based optimization problem, has proved its superiority over variety benchmarking and real time problems. Measuring and visualizing the changes in the diversity of DE population during its search is one of the ways to understand the algorithmic behavior of DE. This helps to provide better insight for proper tuning of control parameters of DE. Hence, an extensive study to describe various possible ways to measure the population diversity of DE algorithm would be a useful tool for the researchers and practitioners of DE. Towards this research direction, this paper presents variety of population diversity measurement methods available for population based algorithm (in general). As well as, as an initial attempt, three methods out of all the identified methods are implemented for DE/rand/1/bin algorithm for a benchmarking function suite with four different functions. The results recorded are presented and discussed in this paper.