INTERVAL METHODS IN KNOWLEDGE REPRESENTATION

In this issues, we continue to publish abstracts and reviews of recents papers on interval methods in knowledge representation. In knowledge representation, intervals are used for two main purposes:

• to describe durations of events; and

• to describe uncertainty of measurement results and expert estimates of different quantities; often, we do not know the exact value of a quantity, but we know its lower and upper bounds (e.g., we may not know the exact value of someone's weight, but we may know that this weight is in between 140 and 160 pounds).

An important case of this uncertainty occurs in knowledge elicitation, when we ask experts to numerically estimate their degrees of belief in their own statements; in this case, it is often difficult for an expert to estimate this degree of belief precisely, but an expert can often provide us with an interval of possible values.

The reviews are collected by Vladik Kreinovich, Department of Computer Science, University of Texas at El Paso, El Paso, TX 79968, USA, email vladik@cs.utep.edu