Skip to main content

(CSC-3734) A Case Study in the Rapid Deployment of New Forecasting Algorithms on a Large Tollway Construction Program

In order to receive proof of CEU credits, you must watch this presentation in its entirety and complete the survey.

Level: Intermediate
TCM Section(s):
6.1. Asset Performance Assessment
Venue: 2021 AACE International Conference & Expo

Abstract: The COVID-19 pandemic resulted in commercial and personal vehicle traffic volumes declining drastically over the course of a few weeks in the spring of 2020.For highway agencies that rely on toll revenue as a substantial part of their funding, this rapid reduction in revenue presented a potential impending liquidity crisis.Agencies with large construction programs underway were forced to balance ongoing cash needs to support those projects with both a drop in revenue and a level of economic uncertainty likely not seen in over a decade.The need for frequent updates of projected progress on large capital programs became critical in order to constantly re-evaluate cash position, changing revenue projections and both short term and long-term impacts.

This paper will present solutions that were developed to provide rapid forecast updates when they could not be produced within the constraints of existing tools and systems that had been configured to measure progress over an annual or quarterly time frame.Suddenly, weekly updates were needed.Using an open source statistical programming language, a methodology was created that allowed for time phased forecasts to be provided to program stakeholders each week on over 300 projects totaling over $1 billion of annual spending.Robust and flexible forecasting algorithms based on evolving information, historical progress data and the latest exploratory data analysis (EDA) were developed and applied to continue successfully managing a $14 billion capital program.