Multi-objective optimisation of an outranking model for public resources allocation on competing projects

This paper develops a model for composing social-oriented project portfolios. The information concerning the quality of the projects is in the form of a project-ranking, which can be obtained by the application of a proper multi-criteria method, but the ranking does not assume an appropriate evaluation of social impact. Inspired by outranking methods, the model provides a preference relation over project portfolios related to project rank positions, single-project costs, and the Decision-Maker's aversion to expensive projects. A best portfolio is primarily found through a multi-objective optimisation that regards violations to pre-established decision-maker's preferences and the competing portfolios' cardinalities. The overall solution is obtained by developing an evolutionary method, which is found to perform very well in some test examples.