In this paper we present a new Ants System approach to a dynamic Travelling Salesman Problem. Here the travel times between the cities are subject to change. To handle this dynamism several ways of adapting the pheromone matrix both locally and globally are considered. We show that the strategy of smoothing pheromone values only in the area containing a change leads to improved results.
|Number of pages||11|
|Publication status||Published - 2002|
- HMI-CI: Computational Intelligence