Capital Budgeting of Interdependent Projects with Fuzziness and Randomness

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 described by means of fuzzy variables. In order to be able to process such hybrid data, a model of the problem is proposed. The model takes into account both stochastic and economic interdependency between projects. Additionally, a new hybrid method for solving this model is developed. The method combines stochastic simulation with arithmetic on interactive fuzzy numbers and nonlinear programming. As a result a set of Pareto-optimal alternatives is obtained. In order to illustrate the performance of the proposed hybrid method, an example from metallurgical industry is provided.