Applications of Multicriteria Portfolio Models in Project Evaluation and Resource Allocation

Multicriteria project evaluation and resource allocation decisions are central and recurrent activities in business and public administration alike. These problems are typically characterized by large number of project proposals, portfolio balance requirements and other constraints; they are also often pressed by urgency and limited data availability. Simple additive scoring models, which o er a transparent structure and are applicable in various problems, have won popularity as the evaluation framework in applications. The formal model needs not be exhaustive, but ease of use combined with facilitated discussions often outweighs more complex technical models. This Thesis synthesizes applications of additive scoring models and develops the Robust Portfolio Modeling (RPM) framework for multicriteria project portfolio analysis. The framework o ers a simple project evaluation structure, allows incomplete information about model parameters and uses linear portfolio optimization to capture multiple resources, project interdependencies and portfolio constraints. Incomplete information implies a dominance concept between portfolios, and consequently leads to novel projectspeci c Core Index, which is based on the computation of all non-dominated portfolios. The Core Index is a robust measure of project's relative t into the portfolio, and it can be used to identify groups of outperformers and underperformers, and projects to be focused on in further analysis and negotiation. Experiences from applications suggest that the key concepts of RPM are readily understandable and deemed useful. Analysis provided by the Core Index works well in translating the problem into actable project-oriented recommendations. Computation of non-dominated portfolios and subsequent interactive analysis are supported by the RPMDecisions c © software. In addition to further developments of the analysis and applications, RPM could be conceptually extended to support hierarchical resource allocation and problems where the portfolio is updated from continuous ow of project proposals. Number of pages: 112