Decision Programming for Solving Optimal Strategies for Multi-Stage Problems under Uncertainty

Event Date: 27 November 2018

Speakers: Ahti Salo, Juho Andelmin and Fabricio Oliveira, Aalto University School of Science, Finland

Location:Strathlcyde Business School, Cathedral Wing, CW404a

Time: 16:00-17:00

Abstract:

Many multi-stage decisions under uncertainty can be structured as influence diagrams which consist of decision, chance and value nodes as well as arcs which represent dependencies between these nodes. Typically, the decision strategy which maximizes the decision maker's (DM) expected utility is determined either by carrying local transformations (such as arc reversals and node removals) or by formulating the equivalent decision tree which is then solved with dynamic programming.

In this paper, we develop an approach called Decision Programming (DP) in which such decision models can be solved by converting them into equivalent linear programming problems. In the context of project portfolio optimization, Decision Programming can be viewed as an extension of Contingent Portfolio Programming (CPP; Gustafsson and Salo, 2005) to problems which involve endogenous uncertainties in the sense that the probabilities of the underlying scenario tree can depend on project decisions. More generally, Decision Programming offers enhanced modelling possibilities in that (i) earlier decisions need not be known when making later ones, (ii) it is possible to introduce both deterministic and chance constraints on the use of resources and (iii) the objective function can be extended to account for risk preferences.We present illustrative examples and provide evidence on computational performance.

References

J. Gustafsson, A. Salo (2005). Contingent Portfolio Programming for the Management of Risky Projects, Operations Research 53/6, 946-956.

Published: 13 November 2018