PLAN⋄HOV: Case-Based Reasoning Planning Tool for High-Occupancy Vehicle Lane Analysis in a Geographic Information System Environment

Abstract

Historical cases can be used to inform new planning and policy situations. Existing knowledge about high-occupancy vehicle (HOV) lane cases can provide quantitative and qualitative information to support decisions about proposed HOV lanes. Using case-based reasoning (CBR) to formalize the use of historical cases, a computer program called PLAN⋄HOV is demonstrated. CBR complements formal modeling by systematically presenting knowledge to users in situations in which formal models are under development or qualitative information is important. PLAN⋄HOV is a planning tool developed in a geographic information system environment that allows users to understand the potential impacts of proposed HOV lanes on the basis of experiences with similar HOV lanes. All historical HOV lane cases represented in PLAN⋄HOV have HOV lane attributes, associated impacts, and the implementation context. In addition, historical cases contain information about lessons learned, such as inferences about their success or failure, prescriptions, and case quality. The historical cases can be accessed by matching them with the dimensions of the current context (e.g., population and network structure) and the proposed HOV lane actions, such as adding a 2+-person HOV lane that operates during peak periods. The retrieval and display of historical cases can occur with varying levels of matching stringency. Users can examine the matched case(s) for their qualitative content (such as controversies over HOV lanes) and also use them to anticipate the impacts of the proposed lane (e.g., mode shift caused by HOV lanes). The development and use of this case-based planning tool are described.