Combining heterogeneous temporal information: A case study

Abstract

In this paper we present a specialized system for temporal reasoning representing possibly uncertain qualitative and quantitative constraints relative to temporal intervals in a homogeneous framework. The underlying logic is that of Allen, which we have extended to include quantitative temporal information. Intervals are still assumed as primitive entities while, to represent metric information, time points are introduced only at the algorithmic level. We illustrate the properties of the temporal representation, the algorithms used for constraint propagation, and their complexity. We show how the system can be used to schedule manufacturing operations in a small factory.