...short introduction...

PPDS is the abbrevation for Project Portfolio Decision Support and is the topic of my thesis which i wrote at the technical university of vienna. I applied a mathematically based approach using mPOEMS to the portfolio project selection problem and embedded this approach into a portfolio process. For the research evaluation test data was created with COCOMO II and the used test set is published on this page in the download section. To get an idea of the work presented in my thesis, the abstract of the thesis is presented below

A poster presenting the topic of my thesis is also available to download.

...abstract...

Multi-Objective Optimisation, Group Decision Support and Visualisation are the major fields of research in this thesis. The optimisation part deals with solving the demanding project selection problem, especially focusing on selecting software development projects. The fields of Group Decision Support and Visualisation are addressed by creating a portfolio selection process, applicable to group decisions, supported by visualisation and data exploration techniques. Students and researchers interested in the field of optimisation, especially with evolutionary algorithms, project portfolio management in general, and group decision support, are the main audience of this thesis. Managers and consultants can find practical and valuable information to improve the project portfolio management of an organisation.
Project Portfolio Management is a set of processes used to support a business in conducting the mix of projects, which best fit the organisation's various needs. This set of processes includes the selection of projects an organisation conducts, maintaining the selected projects in portfolios, and periodically reviewing the mix of projects, to check whether the selection still supports the main business goals. Strategic alignment of the selected projects, efficient use of scarce resources, and balanced portfolios regarding risk and category alignment, are among the main goals of Project Portfolio Management. The selection of projects is a complex task, since numerous goals, restrictions, constraints and subjective preferences have to be taken into account. The complexity grows exponentially with the number of projects to select.
In this thesis a two-phase decision support framework is presented, which focuses on the selection of projects, using an evolutionary optimisation algorithm. In the first phase, an implementation of a mathematically based optimisation framework is presented, using the evolutionary algorithm mPOEMS, based on general goals which the portfolio management should support. This optimisation framework is able to find efficient portfolios on the pareto-optimal front. In the second phase, a decision support process for the selection of one portfolio out of the pareto-optimal front is presented, especially designed to be applicable to a group of decision makers.
A comparison of the presented optimisation approach using mPOEMS, with the two state-of-the-art optimisation algorithms NSGA-2 and SPEA-2, shows that mPOEMS performs better in two out of three multi-criteria optimisation metrics. The main contribution of this thesis is the combination of an objective optimisation project selection framework, with an easy-to-use group decision process, in which individual preferences can be visualised and the search for consent in subjective preferences is facilitated. The presented evolutionary approach to portfolio optimisation is especially meaningful, because a new method about how to deal with the sequencing of projects is outlined. Methods are presented to optimise portfolios with respect to certain objective preferences like category assignment and risk distribution, the consideration of numerous restrictions, constraints and logical dependencies. The presented approach is theoretically not restricted in the number of projects it can deal with.

...some downloads...

• Poster of the PPDS Thesis
• COCOMO II Test Set