The work we describe addresses the problem of deciding between project-funding opportunities under budget and headcount constraints. Although the projects lead to products that yield revenue in the market, complex interactions between the projects and products make the selection of a portfolio difficult. Furthermore, the senior managers in the company have a wealth of business intuition that can inform the required decisions. We combine modeling, simulation, and optimization techniques to provide a set of the best portfolios possible from the proposed projects and resulting products. We also provide a rich set of analysis and visualization tools for the decision makers to use in exploring the suggested portfolios and applying their intuition to make the final selection. The resulting interplay between analytics and intuition produces better business solutions through a more focused and effective debate in a shorter time than previously achieved.

Project portfolio selection (PPS) is a complex problem faced by major companies whenever there are multiple funding opportunities with insufficient budget to fund them all. In this paper, we present our work on a PPS decision support tool that has become a fundamental part of the project portfolio decision process at Intel Corporation across its largest product and market divisions. The paper builds on a previous publication that outlines the decision support tool's bicriteria optimization model by providing a solution procedure that is capable of solving real-life PPS problems within time frames acceptable to decision makers, as well as providing further details on the data collection and the decision-making process. We also report on various analysis and visualization tools that have been built to allow decision makers to interact with promising solutions provided by the decision support tool. One of the contributions of the paper is to present a typology of the important dependencies between projects that needs to be considered, and provide details on how they are incorporated in the decision support tool's optimization engine. We discuss important implementation details on the decision-making process and the agents involved, and conclude by describing real-life experiences on how the framework can enable intuitive decision-making when choosing portfolios that best satisfy the organization's business goals.