Global Sensitivity Analysis in Presence of Correlated Cash Flow

摘要

Sensitivity analysis is often connected with uncertainty analysis. In this context, we use sensitivity analysis techniques for evaluating investment project under uncertainty and risk in presence of cash flow correlations. Variance of cash flow is considered as source of variation in our models. The models estimate two investment criteria: net present value (NPV) and value at any time t (Vt). We first introduce three GSA (Global Sensitivity Analysis) techniques namely, Sobol’ global sensitivity indices, SRC and PEAR respectively but only use one of them because of its superior property to others. We show that if decision maker selects NPV or one of its generalized form as a evaluation criteria: (a) GI (Global Importance) of cash flow can be calculated analytically (b) has a straight forward interpretation in terms of uncertainty management (c) correlation coefficient between cash flows are non negligible. We illustrate our models with a sample in energy sector which exist in literature and compare them with each other. Moreover to confirm presented models sixteen test problems are generated and a non-parametric statistic is carried out. Conclusions and future research directions are provided finally.