ICT develops AI tool for optimal mix design of flexible roadways

4/30/2025 McCall Macomber

An Illinois Center for Transportation team, led by director Imad Al-Qadi, has developed a tool to help contractors, consultants and agencies to develop optimum roadway mix designs by meeting a preset performance criteria or predicting expected performance.

The proposed approach will reduce trial and error in the mix-design process, saving time and costs, by cutting down the design period from two to three weeks to within a day.

Before hitting roadways, asphalt mixes in Illinois must pass potential performance limits for cracking susceptibility and permanent deformation potential, which are evaluated using the Illinois Flexibility Index Test, developed by ICT, and Hamburg Wheel-Tracking Test, respectively.

To avoid pay disincentives during construction, Illinois contractors must find the correct proportion of materials when designing asphalt mixes to pass the tests’ required limits — or risk needing adjustments to the mix and restarting the design process.

The ICT tool harnesses machine learning and genetic algorithms — forms of artificial intelligence — to learn from extensive state data provided by Illinois Department of Transportation to improve the algorithms’ accuracy.

“This tool uses a data-driven method that significantly minimizes or eliminates the trial-and-error process required for HMA [hot-mix asphalt] design, providing transportation agencies and consultants with a cost-effective, performance-based solution to improve pavement quality and durability,” said Hong Lang, ICT postdoctoral research associate.

A screenshot of the ICT optimal mix design tool. The tool features three manuals that allow users to gain a general understanding of the tool, run two prediction modes for optimal mix design and conduct a volumetrics check.
A screenshot of the ICT optimal mix design tool. The tool features three manuals that allow users to gain a general understanding of the tool, run two prediction modes for optimal mix design and conduct a volumetrics check.

The open-access, web-based tool, which relies on a performance-based approach, allows users to input their own aggregate data using two prediction modes as well as supports volumetrics checks for actual test analysis.

Users may run the tool to quickly predict either mix design for a desired flexibility index and rutting depth or predict those values for a known design.

“Automating the mix design through calculating an optimized asphalt binder content and aggregate gradation to meet (or surpass) required performance limits would reduce time, costs and material required in the mix-design process,” said Imad Al-Qadi, ICT director and Grainger Distinguished Chair in Engineering. “The impact on the flexible pavement industry in Illinois is expected to be significant.”

Contributions to the tool are as follows: concept and design (Imad Al-Qadi), data analysis and interpretation of results (Hong Lang, Al-Qadi and Uthman Mohamed Ali), and platform programming and maintenance (Lang).

Learn More