Handling of Alternatives and Events in Temporal Databases

Abstract

Planning for the future is an important activity both at the individual and organizational levels. Planning consists of defining alternative actions to handle various events in the future. The alternatives arise becau]se of different possible outcomes of events. A plan consists of a sequence of actions to be carried out for each possible outcome. In the context of database modeling, the actions are operations on a database. A database management system should enable its users to define events and alternatives, and also allow them to interact with the database under different alternatives (possibly to evaluate different plans). The existing temporal data models treat the future analogous to the past or present; they provide for one future path (in the sense that facts valid at some future time can be stored), but do not provide support for alternatives in the future. In this paper, we present a model for incorporating events and alternatives by extending the temporal data model to support branching time. The extended model permits definitions of events, their interdependencies and associated actions. The events that affect an object are modeled by a tree, permitting an object to have different states at the same valid time but under different alternatives. The branching time paradigm is obtained by superimposing a linear valid time on the event tree. We extend the temporal relational algebra and the Temporal SQL2 to support a branching time data model. The paper also briefly deals with the uncertainties associated with future planning as well as probabilities of possible event outcomes. Finally, we sketch an implementation strategy for the branching time data model.