Differential Evolution (DE)
Erik Cuevas, A. Gómez Rodríguez · 2020
In 1997, Kenneth Price and Rainer Storn proposed the metaheuristic algorithm known as Differential Evolution (DE). This approach is a population algorithm that implements a direct and simple search strategy. Under its operation, DE considers the generation of parameter vectors based on the addition of the weighted difference between two members of the population. In this chapter, the operative details of the DE algorithm are discussed. The implementation of DE in MATLAB ® is also described. The objective of this chapter is to provide the reader of the mathematical description of the DE operators and the capacity to apply this algorithm in the solution of optimization problems. To do this, in the subsequent sections, the use of the DE algorithm is considered in two aspects: the first is the resolution of an optimization problem minimizing a mathematical benchmark function, and the second, the solution of an engineering problem that requires an optimal design in their parameters considering some design restrictions.