Using a branch and bound method to solve realistic TSP

An algorithm to solve TSP based on a branch and bound thought is presented, and the calculation of upper and lower bounds is discussed in detail. The calculation of lower bound's value is based on the current path, it is simple and easy and take up less memory. The upper bound has only a global value. One of the characteristics of realistic TSP-triangle inequality is used in the calculation, the result does not exceed 1.5 times the best value. Another characteristic of realistic TSP is symmetry, it can narrow the tree solution space a half, and further speed up the search process. The algorithms of seeking lower bound and upper bound is independent and can fully separated from each other, but through the example, it is proved more effective of combining the two method with the thought of branch and bound, and it can greatly accelerate the search of the tree of solution space.