| Article ID: | iaor1997438 |
| Country: | United States |
| Volume: | 43 |
| Issue: | 5 |
| Start Page Number: | 749 |
| End Page Number: | 763 |
| Publication Date: | Aug 1996 |
| Journal: | Naval Research Logistics |
| Authors: | Erkut E., Zhang J. |
| Keywords: | programming: travelling salesman, programming: branch and bound, heuristics, networks: path |
The authors consider a routing problem where the objective is to maximize the sum of the rewards collected at the nodes visited. Node rewards are decreasing linear functions of time. Time is spent when traveling between pairs of nodes, and while visiting the nodes. They propose a penalty-based greedy (heuristic) algorithm and a branch-and-bound (optimal) algorithm for this problem. The heuristic is very effective in obtaining good solutions. The authors can solve problems with up to 20 nodes optimally on a microcomputer using the branch-and-bound algorithm. They report our computational experience with this problem.