Last modified: 2014-01-23
The Traveling Salesman Problem (TSP) is one of the well-knows optimization problems. The importance follows from many practical applications and from its computational complexity as well. It is possible to use many heuristic techniques that lead to suboptimal solution for TSP. Presented algorithm (Differential Evolution) belongs to Evolutionary Algorithms that were originally focused on solving mathematical programming tasks containing continuous variables; therefore the use of Evolutionary Algorithm for combinatorial problems requires some special approaches. In this article the use of Differential Method will be presented for solving TSP (8 cities set and 25 cities set). Firstly, the penalty approach is applied as a simple way to guarantee feasibility of solution. Then, new approach that works only on feasible solutions is presented.