Dealing with time in diagnostic reasoning based on causal models

Temporal reasoning has not received so far in the field of diagnostic problem solving the attention it deserves. However, the notion of time is strictly connected with the notion of causality whose role in diagnostic reasoning is rapidly growing. In this paper we propose an approach to represent temporal knowledge within deep causal models and we suggest a strategy to combine causal and temporal reasoning for diagnostic problem solving. More specifically, a scheme to perform temporal reasoning on a causal model and some criteria to define the temporal consistency of a causal explanation are described in full detail. Particular attention is paid to the analysis of the temporal reasoning component and to the computational complexity of the temporal reasoning scheme we have defined.