The ideas of a 19th-century Russian mathematician are helping BorgWarner Inc. minimize costly parts disruptions and other glitches in its supplier network.

BorgWarner says its automated supplier-monitoring system, coming soon to the United States, can identify a looming supplier problem so early and accurately that it sometimes flags the issue before managers at the supplier do.

The company's Supplier Performance Monitor, which it began using in Germany and Ireland about two years ago, has saved money by reducing parts defects and allowing BorgWarner to stock fewer surplus parts.

Holger Kiebel, vice president of supply chain at BorgWarner unit Beru Systems in Ludwigsburg, Germany, says both metrics have declined by double-digit percentages.

About 180 suppliers are being monitored. BorgWarner plans to implement the system in the United States and other parts of Europe this year and in India and China next year, Kiebel says.

He says the system mainly analyzes the quantity of parts in each delivery, comparing that figure with the original order and with the number of parts that have to be returned for quality problems.

It analyzes each delivery on a much more detailed level than the company was able to do before, which helps estimate the potential for future problems. An order might be flagged, for example, if a batch is supposed to be delivered all at once but the parts come in multiple shipments that deviate from the schedule.

Kiebel says about 85 percent of the discrepancies identified by the system have proved accurate. Nearly all of the rest have stemmed from inaccurate delivery data.

When a potential problem is spotted, employees at BorgWarner contact the supplier to confirm it, offer help if needed and begin working on a solution as quickly as possible.

"It's a completely different kind of communication because we can discuss it before something has happened," Kiebel says. "There's enough time to take the right action, and suppliers really like this kind of cooperation."

He adds: "One supplier called back and said, 'You were right, and I was not aware of it until you called.'"

Dave Andrea, senior vice president for industry analysis and economics at the Original Equipment Suppliers Association, says rapid growth in the auto industry means that even minor breakdowns in the supply chain can quickly have far-reaching consequences.

He says suppliers are operating at about 80 percent of their production capacity on average, with a quarter of them at 95 percent or higher and many at more than 100 percent, leaving little room for disruptions.

"We've gone from a time where we were managing the excesses to one now where we really are managing the constraints," Andrea says. "It takes more attention in terms of watching for troubled suppliers and looking for red flags."

Automakers and suppliers go to great lengths to monitor their direct suppliers but have a hard time obtaining useful insight into those companies' suppliers. Increasingly, Andrea says, companies are sending quality-assurance teams to several tiers of suppliers to check on inventories, maintenance and other variables that could hinder production.

A record number of new-vehicle launches also has upped the level of pressure and complexity that many suppliers must deal with, he says.

The BorgWarner system is based on the Hidden Markov model, a concept researched by mathematician Andrey Markov in the late 1800s.

The system assigns each supplier a three-letter code, ranging from AAA to CCC. The first letter represents that company's current status, the second is its outlook for the next several weeks and the third is its medium-term forecast.

"To know today what a supplier delivered yesterday is too late," Kiebel says. "We want to know what a supplier will be delivering the day after tomorrow."