Interval relations play a significant role in constraint-based temporal reasoning, resource allocation and scheduling problems. For example, the intervals may represent events in time which may conflict or may be compatible, or they may represent tasks to be performed according to a timetable which must be assigned distinct resources like processors or people. In previous work [G93, GS93, G98], we explored the interaction between the interval algebras studied in artificial intelligence and the interval graphs and orders studied in combinatorial mathematics, extending results in both disciplines. In this paper, we investigate algorithmic problems on tolerance graphs, a family which generalizes interval graphs, and which therefore have broader application. Tolerance graph models can represent qualitative and quantitative relations in situations where the intervals can tolerate a certain degree of overlap without being in conflict. We present a coloring algorithm for a tolerance graph on n vertices whose running time is O(n), given the tolerance representation, thus improving previously known results. The coloring problem on intervals has direct application to resource allocation and scheduling temporal processes.