ABSTRACT
A certain ambiguity over which is the better scheme now arises because the distributions of their associated returns overlap. The decision must revolve around the trade-off between a mean expectation of high returns or low costs and the greater security of a low variance about the expected value. Cole presents a procedure for the treatment of uncertainty at the project design stage of planning where it is legitimate to construct subjective probability distributions of the performance of a number of discrete designs for a facility. Roberts presents a programming formulation for the maximisation of expected utility, measured by some cardinal index. The arguments of this utility function and the benefits of possible projects are stochastic and capable of representation by subjective probability distributions. The formulations reviewed thus far require knowledge sufficient to compute the probability of various configurations of demand and performance associated with particular courses of action.
