Contractor selection based on predictive performance

Abstract

This paper proposes a contractor selection system that incorporates the contractor performance’s prediction as one of the criteria for selection. This research was developed working with an owner organization that was interested in developing a framework for evaluating contractors for future works. A modeling framework, developed in previous research, was used to develop a conceptual model of a project that depicts a causal structure of the variables, risks and interactions that affect to contractor’s performance for a specific project from the owner’s point of view. The conceptual model helps to identify the needs of information for a comprehensive evaluation, some information can be readily available from historical records while other can be unavailable and can be replaced by estimates based on experience. Ideally, over time the owner should collect most of the information required for future evaluations. A mathematical component of the model can generate predictions of multiple project performance outcomes for each contractor under evaluation; these predictions, joined with the contractors’ bid prices, are then used for contractor evaluation purposes.