Hybrid differential evolution with the topographical heuristic

In this article, we present a new hybrid differential evolution (DE) which employs a topographical heuristic introduced in the early nineties as part of a global optimization method. This heuristic is used to select individuals from the DE population in order to be starting points of instances of the Hooke–Jeeves algorithm. The solutions achieved in this phase are potential candidates for the next generation. The method, called TopoDE, is compared with other stochastic optimization algorithms using challenging benchmark problems. The results obtained are quite promising.