Developing a Virtual Laboratory for Construction Bidding Environment Using Agent-Based Modeling

Abstract

The competitive construction bidding process is a highly unstructured decision-making problem where contractors have to take several factors into consideration, weigh their relative importance, and consequently make bidding decisions within the constraints of bounded information, limited time, and uncertainty about the market constituents. The bidding models presented in the literature help identify the different bidding objectives a contractor would consider but fail to capture all the complexities of this mechanism and its dynamics. This paper presents an agent-based construction bidding model that allows modeling different bidding strategies of contractors, observing their interaction and learning over time, and detecting emergent market behavior and bidding patterns. This model is used to simulate a variety of bidders competing against each other under a low-bid tendering approach in order to study the effect of different bidding parameters on their optimal markup decision, long-term financial performance, and survival in the market. Specifically, this paper observes the effect of the market competition level, a contractor’s risk attitude, his cost estimation and project management skills, and the project complexity on the contractor’s bidding decisions and market equilibrium.