This paper discusses the problem of capital budgeting in the situation where only some model parameters are well described by their past data and therefore are specified by random variables, whereas the remaining model parameters can hardly be predicted by historical data and therefore they are...