Planning and temporal reasoning under uncertainty

Abstract

This paper describes a mechanism to assist in robot problem solving which handles queries concerning temporal knowledge involving uncertainty. The program or set of programs, which I will refer to as a task network management system or TNMS, is capable of detecting and responding appropriately when assumptions made in the course of problem solving are violated during subsequent planning. The paper focusses upon the functionality of processing queries and the problem of coping with the computational complexity interent in dealing with time.