摘要: |
Case-based reasoning (CBR), an emerging artificial intelligence (AI) technique, offers the potential to provide for the real-time functionality required of a real-time traffic flow management decision support system. CBR solves new problems by reusing solutions of similar past problems. It is based on the observation that when people solve a new problem, they often base the solution on one that worked for a similar problem in the past. The motivation behind adopting a CBR approach for real-time traffic routing lies in the fact that a CBR system, by reusing successfully routing strategies for similar conditions from its case-base, will avoid the need to solve the problem from scratch each time. The concern, however, with using CBR for a problem such as real-time routing is that the components of the problem might force the need for prohibitively large case bases in order to achieve satisfactory performance. It thus becomes particularly important to design a case-base that adequately covers the range of problems the system is expected to face, and at the same time, to keep its size manageable. In this study, a case-base satisfying these two requirements is designed for a prototype CBR routing system. The system is developed for a real-world highway network in the Hampton Roads region of Virginia. Cases for building the system's case-base are generated using a heuristic DTA model specifically developed for the region. The adequacy of the case-base design is tested by comparing the solutions generated using the CBR approach and the DTA model solutions. The results indicate that the recommended design will allow a CBR system to function in real-time and to produce high quality routing strategies. |