Uncertain temporal knowledge management

This paper presents an approach of managing temporal knowledge under uncertainty based on fuzzy logic and possibility theory. Temporal uncertainties or modelled in terms of possibility distributions. A Fuzzy Timed Causal Like Network (FTCLN) is defined to represent knowledge about the process or system under consideration. A FTCLN is, in fact, a timed acyclic directed graph where a node is the description of an event and an arc the temporal relation between the two events it links. All events are dated and are ordered according to the moment of time when they occur. The direction of the arc indicates the qualitative temporal order (symbolic aspect) and the length of arc describes the time interval between two events. The temporal reasoning under uncertainty is performed by evaluating this network and a prototype has been implemented which leads to the F-TRYST (fuzzy temporal reasoning system).<>