MDT is required by federal regulations to have a procedure and tools for contract time determination. The contract time drives much of the contractor’s approach to bidding a specific project. If the MDT sets a period that is shorter than the construction industry’s estimate, bids will go up across the board as competing contractors bid the cost to accelerate the project. If the period is longer than the industry’s estimate of time, the bids also go up as contractors bid the additional costs to remain mobilized. Either way, the public loses due to inaccurate and sometimes arbitrary contract completion times. Emerging Artificial Intelligence (AI) algorithms have power to process various types of data and learn the hidden patterns and make a reasonable prediction with reliable accuracy. This research project will use one of the promising AI algorithms, namely, the neural network algorithm to quickly estimate the most likely contract time for a highway project. The principle idea is to use key project characteristics that may include work type, project location, major controlling work items and their quantities of work as input variables and then, the neural network algorithm will estimate the most likely project duration or contract time of a new project by analyzing the data of the historical MDT highway projects. One of the major challenges that MDT schedulers face is the short period of time allowed for contract time determination. Thus, a quick and effective contract time determination tool that produces high-quality results in a more reliable and defensible manner will serve as a great supporting tool for improving work efficiency.

Project Overview
Proposal
Kick-Off Meeting Notes
Task 1 Report

For more information, contact Vaneza Callejas, 802-546-0217.