Handling incompleteness in reasoning: application to situational analysis and real time reactive systems

There are many situations where a human or machine reasoner must make inferences even when the information available is incomplete. Various approaches have been proposed regarding which inferences should be made under these circumstances. Most of the proposals for extending incomplete theories (closed world assumption, circumscription and default theories) choose a single preferred extension. However in many cases, reasoning about possible and relevant worlds is required. The authors propose a new approach to handling incompleteness called constraint quantification which generalizes both existential and universal quantification. The use of this method to analyze contingencies in situations and the application to reactive systems for real-time reasoning is discussed