Case-based reasoning: A planning tool for intelligent transportation systems

Abstract This paper develops a Case-Based Reasoning (CBR) methodology for PLANiTS (Planning and Analysis Integration for Intelligent Transportation Systems). To address a current transportation planning situation, CBR presents similar historical cases. Specifically, it estimates the impacts of proposed transportation improvement actions, including Intelligent Transportation Systems, based on previous experiences with similar actions. In this paper, a hierarchical structure for representing historical cases is developed. All historical cases consist of transportation improvement actions, performance measures and environments defined in terms of their spatial, temporal and user/traveler dimensions. In addition, cases contain information about lessons learnt such as inferences regarding their success or failure, prescriptions and case quality. Information about historical cases can be synthesized. Specifically, the reasoner contains mechanisms for 1. (a) matching at various levels of stringency, 2. (b) ranking with alternative distance and weight measures 3. (c) analyzing similar past cases with statistical operations. Also discussed are the limitations of CBR applications to transportation planning. Overall, the structure for the CBR is flexible and incorporates different stakeholder preferences for alternative transportation improvement actions and evaluation criteria. Further, the application of CBR to transportation planning, discussed in this paper, formalizes the use of case knowledge in transportation. This can lead to improving the transportation planning process.