If urban development plans were just target patterns to be achieved, conventional data structures in Geographic information systems (GIS) would be sufficient. Urban development plans have a strong spatial component, but recent literature in planning emphasizes that plans are about actions and relationships among them. These relationships include interdependence, substitutability, priority, and parthood. In order to support planning, GIScience should devise data structures and queries to support reasoning with these relationships. This article shows how relationships encoded within each of a set of plans, using a recently developed data model, can be used to infer the relationships of actions among these plans. Simple databases and use cases based on real situations in McHenry County, Illinois are used to demonstrate that these relationships can be encoded and queried. The results demonstrate that previously discovered semantic relationships can be used to discover additional relationships across plans, thereby enriching the decision making. The approach provides a systematic way of structuring the information in plans to support making and using plans.
IntroductionConventional geographic information systems (GIS) cannot support reasoning with plans: How does an action affect previous actions, actions of others, and actions that come after? How can we embed a decision to commit to a particular action in the reasoning already embedded in plans? How can we use the information available at the time of action, including the information in plans, to decide what to do? How can we reason efficiently, but still imperfectly and incompletely, using any information made newly salient by our particular focus of attention? Recently articulated conceptions of plans in GIS and GIScience (e.g. Couclelis 2009), artificial intelligence (AI) (e.g. Pollack and Horty 1999), and urban development (e.g. Hopkins 2001) converge in suggesting that spatial representation is insufficient and that substitutability with respect to intentions and effects and interdependence among actions are crucial relationships that need to be captured in information systems supporting planning. By representing these actions only as spatial entities, we lose the underlying logic that binds these actions together. This article is an attempt to augment spatial reasoning with plan-based reasoning in response to the challenge from Couclelis (1991).