Abstract

CANDIA-GARCIA, Cristian David; LOPEZ-CASTRO, Luis Francisco  and  JAIMES-SUAREZ, Sonia Alexandra. Optimal Project Portfolio Selection Using Meta-Optimized Population and Trajectory-Based Metaheuristics. Rev.EIA.Esc.Ing.Antioq [online]. 2020, vol.17, n.34, pp.271-288.  Epub Sep 06, 2021. ISSN 1794-1237.  https://doi.org/10.24050/reia.v17i34.1399.

This article addresses the problem of project portfolio selection for the awarding of public works audits through open merit competitions (CMA) supervised by the National Roads Institute in Colombia - INVIAS. In this modality, each competitor presents a unique portfolio of historical projects to quantify its experience. As an alternative to the use of Excel spreadsheets with limited procedures of exhaustive enumeration, a meta-optimized genetic algorithm (GA) and a meta-optimized greedy randomized adaptive search procedure (GRASP) were evaluated for the case study of a company with 207 experience career contracts. Both metaheuristics were able to find optimal assessment scores for different test instances, however, the GA algorithm consistently performed better in all assessment instances, finding in some cases up to 10 optimal portfolios in less than 9 minutes.

Keywords : genetic algorithms; GRASP; meta-optimization; project portfolio selection.

        · abstract in Spanish     · text in Spanish     · Spanish ( pdf )