Static R&D project portfolio selection in public organizations

• A multi-objective model framework where projects are sets of tasks is proposed.
• The model framework includes synergies between tasks or projects.
• The presented model framework supports both partial and total funding policies.
• Instances ranging from small to large are solved efficiently in a standard solver.
• Factors affecting solution quality and computational load are statistically examined.

The problem of static research and development (R&D) project portfolio selection arises when a public organization opens a call for proposals and then builds a portfolio choosing which ones to fund in terms of impact measures including social objectives, emphasis areas, geographical influence, and other non-monetary factors. The funds are classified into types of expenses, depending on the goals and preferences of the organization: equipment purchases, travel expenses, scholarships, publication fees, etc. We propose a mathematical model framework in which each project proposal comprised tasks with a specific type of expense and the assigned funding may be a fraction of the requested amount, accounting for possible inter- and intra-proposal dependencies. We present computational experiments that show our models to be efficiently resolvable.