Article ID: | iaor20164705 |
Volume: | 63 |
Issue: | 5 |
Start Page Number: | 1117 |
End Page Number: | 1130 |
Publication Date: | Oct 2015 |
Journal: | Operations Research |
Authors: | Sim Melvyn, Chen Lucy Gongtao, Long Daniel Zhuoyu |
Keywords: | programming: dynamic, investment, combinatorial optimization, financial |
We investigate a dynamic decision model that facilitates a target‐oriented decision maker in regulating her risky consumption based on her desired target consumption level in every period in a finite planning horizon. We focus on dynamic operational decision problems of a firm where risky cash flows are being resolved over time. The firm can finance consumption by borrowing or saving to attain prescribed consumption targets over time. To evaluate the ability of the consumption in meeting respective targets, we propose the consumption shortfall risk (CSR) criterion, which has salient properties of attainment content, starvation aversion, subadditivity, and positive homogeneity. We show that if borrowing and saving are unrestricted and their interest rates are common, the optimal policy that minimizes the CSR criterion is to finance consumption at the target level for all periods except the last. For general convex dynamic decision problems, the optimal policies correspond to those that maximize an additive expected utility, in which the underlying utility functions are concave and increasing. Despite the interesting properties, this approach violates the principle of normative utility theory and we discuss the limitations of our target‐oriented decision model.