Predicting the Level of Competition and Determining Optimal Bidding Strategies for Bundled Projects: Integrating Machine-Learning Algorithms and Probabilistic Modeling

Project bundling involves grouping multiple projects into one contract to improve efficiency and reduce costs, especially in transportation infrastructure. Contractor bidding behavior differs on bundled contracts compared to single-project contracts due to various unique risks and complexities inherent in bundled contracts. Given the lack of existing guidance in assisting contractors in their bid preparation on bundled contracts, this paper develops a competitive bidding framework to help contractors in improving their pricing strategies and bidding tactics on bundled contracts through a series of data-driven models developed on a data set of 1,650 bundled projects. This study developed and compared the performance of various machine-learning models to predict the level of competition (i.e., the number of potential bidders) on bundled contracts based on various input variables about the characteristics of bundled contracts. Statistical distributions were then fitted to estimate the various probability functions of the submitted markup values of all potential competitors, which are useful in modeling the bidding behaviors of contractors on bundled projects. The winning probability and the expected profit for different markup values were then quantified by extending Friedman's and Gates's probabilistic models to the bundling context along with conducting Monte Carlo simulations. Results provided unique quantitative insights on the interplay between the characteristics of bundled contracts, the expected level of competition, the probability of winning, and the expected profit when deciding what markup values to use on bundled contracts. Insights were also obtained from industry practitioners who highlighted that the developed framework is not only theoretically sound but also practically viable. This paper adds to the body of knowledge by developing a new analytical, data-driven framework that not only helps contractors navigate the pricing strategy and bidding landscape of bundled contracts, but also enables public agencies to structure more competitive bundled contracts based on the expected competition.