|
Abstract : |
Abstract. This paper discusses an implementation of a dynamic programming approach to the traveling salesman problem that runs in time linear in the number of cities. Optimality can be guaranteed when precedence constraints of a certain type are present, and many problems involving time windows fall into this class. Perhaps the most interesting feature of the procedure is that an auxiliary structure is built before any particular problem instance is known, reducing the computational effort required to solve a given problem instance to a fraction of what it would be without such a structure. 1, |
