A hybrid Differential Evolution method and its application to the physical travelling salesman problem

Differential Evolution (DE) is a simple and efficient evolutionary algorithm for optimization problems over continuous space. A variant of DE is the Down-hill Simplex method based on Differential Evolution (DSM DE) which has the advantage of converging faster than DE. However, the problem with DSM DE is that it doesn't guarantee to converge to a global optimum. In this paper, we present a way to improve DE by combining DE with DSM DE and the application of the new method to the problem of finding the optimum path in the physical travelling salesman problem.