Organizations must take decisions on how to allocate resources to 'go/no-go' projects to maximize the value of their project portfolio. Often these decisions are complicated by several value criteria, multiple resource types and exogenous uncertainties that influence the projects' values. Especially when the number of projects is large, the efficiency of the resource allocation and the quality of the decision making process are likely to benefit from systematic use of portfolio decision analysis. This Dissertation develops and applies novel methods to manage uncertainty in decision analytic models for project portfolio selection. These methods capture incomplete information through sets of feasible model parameter values and use dominance relations to compare portfolios. Based on the computation of all non-dominated portfolios, these methods identify i) robust portfolios that perform well across the range of feasible parameter values and ii) projects that should surely be selected or rejected in the light of the incomplete information. These methods have several implications for project portfolio decision support. Explicit consideration of incomplete information contributes to the reliability of analysis, which is likely to increase the use of portfolio decision analysis in new contexts. Furthermore, cost and time savings in data elicitation may be achieved, because these methods can give robust decision recommendations based on incomplete data and identify projects for which additional information is beneficial. Finally, these methods support consensus building within organizations as different views about projects' quality or exogenous uncertainties can be considered simultaneously to identify projects on which further negotiations should be focused.
展开▼
