Edges Elimination for Traveling Salesman Problem Based on Frequency \(K_5\)s
We eliminate the useless edges for traveling salesman problem (TSP) based on frequency \(K_5\)s. A frequency \(K_5\) is computed with ten optimal five-vertex paths with given endpoints in a corresponding \(K_5\) in \(K_n\). A binomial distribution model is built based on frequency \(K_5\)s. As the frequency of each edge is computed with N frequency \(K_5\)s, the binomial distribution demonstrates that the frequency of an optimal Hamiltonian cycle edge is bigger than 4N on average. Thus, one can eliminate the edges with frequency below 4N to reduce the number of concerned edges for resolving TSP. A heuristic algorithm is given to eliminate the useless edges. After many useless edges are cut, the computation time of algorithms for TSP will be considerably reduced.
KeywordsTraveling salesman problem Frequency \(K_5\) Binomial distribution Heuristic algorithm
- 1.Johnson, D.S., McGeoch, L.-A.: The Traveling Salesman Problem and its Variations, Combinatorial Optimization. 1st edn. Springer Press, London (2004)Google Scholar
- 7.Thomas, H.-C., Charles, E.-L., Ronald, L.-R., Clifford, S.: Introduction to Algorithm, 2nd edn. China Machine Press, Beijing (2006)Google Scholar
- 8.Mömke, T., Svensson, O.: Approximating graphic TSP by matchings. In: FOCS 2011, pp. 560–569. IEEE, NY (2011)Google Scholar
- 14.Gharan, S.-O., Saberi, A.: The asymmetric traveling salesman problem on graphs with bounded genus. In: SODA 2011, pp. 23–25. ACM (2011)Google Scholar