Decision-making through a fuzzy hybrid AI system for selection of a third-party operations and maintenance provider

Abstract

With the outsourcing and the increasing demand of facilities management services, we observe the growing of multi-technical contracts in real estate operations and maintenance (O&M). Selection of one or more contractors is actually complex and important financial and quality of service challenges depend on it. The present paper proposes a multiple-criteria decision-making tool whose objective is to predict contractors' performances and to select the one who can best respond to O&M demands. In order to build the heuristic between technical, commercial and quality criteria and the expected performances, a neuro-fuzzy system (NFS) associated with a hybrid and adaptive genetic algorithms (GA) method has been developed. Important problems are considered: data pre-processing, problem of data scarcity to provide a sufficient number of data to the NFS and optimisation of hybridisation or adaptation parameters for GA. A case study, concerning the clients' satisfaction levels for O&M contractors as a final indicator for decision-making will prove the relevance of this approach.