The Performance of Ant System in Solving Multi Traveling Salesmen Problem

Multi Traveling Salesmen Problem (mTSP) is an extended situation of Traveling Salesman Problem (TSP) that is classified to combinatorial optimization problems. For many years, TSP has been employed to test the proposed algorithms that belong to meta heuristic techniques. This work is directed to study the performance of ant system (AS) to solve mTSP. This technique was simulated to determine a total shortest route for 5 to 10 salesmen who have to visit 10, 15, 20, 25, and 30 cities. The performance of AS is measured by observing the minimum distance resulted and the average completion time. Ten trials are made for every combination of salesman-city. The result showed that the minimum distance and the processing time of the AS increase consistently whenever the number of cities to visit increase. In addition, a different number of sales who visited certain number of cities proved significantly affect the running time of AS, but did not prove significantly affect the minimum distance.