Pekka Mild Multicriteria Portfolio Analysis for Strategic Resource Allocation

This thesis is concerned with the problem of choosing a subset of projects, a project portfolio, from a large set of multicriteria proposals subject to scarce resources. The overall value of each project is modeled through an additive weighting model, and the value of a portfolio is the normalized sum of its constituent projects ’ values. Various strategic and logical constraints are admitted, as well as project interactions. By the methods of preference programming, incomplete information about the criterion weights of the additive value model can be elicited, thus building up the decision maker’s commitment to the decision recommendations. The thesis extends the use of preference programming to project portfolio selection. A new algorithm for determining the potentially optimal portfolios and their respective optimal weight regions is presented. Recursive use of this algorithm for determining all non-dominated portfolios is also outlined. Performance measures are developed to describe the attractiveness and sensitivity of the different potentially optimal portfolios and individual project proposals. The novel project-specific measures help in the screening of large sets of proposals and convey which projects should incontestably be included in or excluded from the portfolio in view of the incomplete weight information