Allocation of contingency in activity duration networks

This paper describes a method for apportioning contingency in a probabilistic network. It distinguishes between two requirements for contingency: for high variance crucial activities, and for high float noncritical activities. This enables a logical two-step apportionment, the first based upon the idea of 'cruciality', which has become established as an important indicator of an activity's importance in risk terms, and the second based upon standard ideas about float. A numerical example is given based on data used by previous work, so that a comparison of the methods can be made.