A linear approach for solving stochastic capital budgeting problems

Abstract

We consider capital budgeting problems with linear decision variables that can be either continuous or integer and where some or all of the associated cash flows are random variables that may be statistically dependent. They are formulated as convex chance-constrained programming problems that can be approximated by ordinary (integer or noninteger) linear programming problems. The proposed procedure allows the explicit consideration of decision opportunities dealing with a deficit or surplus in periodic net cash flows such as the accumulation of an optimal cash reserve or appropriate borrowing and lending opportunities and of penalties that have to be paid if periodic deficits do occur or the probability constraints do happen to be violated.