An Ant Colony Algorithm for Text Clustering

In this article, a new PMACO algorithm is proposed for capacitated P-Median Problem (CPMP). It devised a set of performing strategies of Ant Colony Algorithm (ACO) in view of the characteristic of CPMP. These strategies include the selecting strategy of initial medians, the pheromone learning strategy of object-assignment means and pheromone-smoothness strategy. They insure the PMACO algorithm can solve the CPMP effectively, which was proved by the computational results. The comparability of Text Clustering Problem (TCP) and CPMP inspired us to employ the PMACO algorithm solving the TCP. As a result, we used the improved PMACO to solve the TCP and obtained preferable result. This exploited the new method of using the ACO to settle TCP self-adaptively.