A search heuristic for the routing of data communication networks
Shih-Yuan Tsen · 1992
In this thesis an algorithm is developed to address the routing problems of data communication networks. The proposed method integrates the heuristic method developed by Frank and Chou [1] for the multicommodity flow model and the hill climbing local search approach. Their heuristic provides an advantage of fast network analysis and is extremely useful during the network design process, however, the performance deteriorates when the heuristic handles larger networks. In order to increase the flexibility and the performance of the heuristic, we adopted a search procedure developed by Storer, Wu and Vacarri [12] to search the neighboring solution space of the heuristic. Results show that the performance of the heuristic increases in general, and particularly for large size networks. The factors that affect our proposed method are investigated by experimental results. A comparison of the proposed method and the base heuristic was made by testing a number of example networks. Detailed description of the methods, including the pseudo codes for both the base heuristic and our proposed method, are provided in this report.