Interval Processing for Search Control Under Incomplete Data

In searching under incomplete information, the utilities or probabilities of the alternative candidates and branches may be available only in intervals instead of precise numbers. Thus interval comparison becomes a key point in search control. This paper introduces, reviews and examines the techniques which compare the intervals exploiting the available information. We also discuss when these techniques can and should be used and give an example of applying a specialized interval ranking mechanism for a scheduling problem. Through this paper we would like to link the interval computation and especially interval comparison into the search community as a considerable issue and helpful techniques for search control under uncertainty and incomplete information.