In the days after the 9/11 attacks, FedEx Ground quickly realized it faced a potentially costly problem: How would it get packages out if even one of its myriad distribution hubs went down, cut off by acts or potential acts of terrorism? The company couldn't just shut down a hub and then see what happened to test the impact.

	Daniel Marsula, Post-Gazette Click photo for larger image.

So a FedEx team studying the issue adapted a computer model to show packages flowing through the network and simulated shutting down different sites. That allowed it to come up with plans for the real world, said John Zicha, managing director of revenue forecast analysis for the Moon-based division

The project used algorithms, a word that for many brings up dim memories of high school math. Merriam-Webster defines an algorithm as a step-by-step procedure for solving a problem or accomplishing some end, especially by a computer. Sounds pretty dull. But the right algorithm can save a company such as FedEx tens of millions of dollars.

The fact is, algorithms are used all over the business world -- and in your world, too. That list of Outer Banks summer rentals that popped up during your online search last night? Or the Amazon recommendation on books that you might like? Or the Pandora music program that helps develop a stream of songs tailored to your specific taste? All are the result of algorithms.

Computer scientists use these mathematical recipes to develop ways to power the Google search engine, protect financial information online, determine which customers respond to coupons and even suggest when chain stores should put merchandise on sale.

Human beings use algorithms, too. Say it's the weekend and you've got errands to do. You make a list, rank priorities based on key factors and try to group trips by location for efficiency.

"When you're thinking about running a lot of errands, you're essentially running an algorithm," said R. Ravi, associate dean of intellectual strategy at Carnegie Mellon University's Tepper School of Business.

But the power of such sorting and looking for patterns to enable decision-making has been magnified as the world's increasingly sophisticated computers pull in piles of data that no human being could ever hope to effectively sort in a useful way.

Multibillion-dollar businesses have been built on having the right algorithm, and companies have become enamored of the possibilities in finding better ones to squeeze savings out of warehouse systems, employee scheduling and fuel costs. Things have progressed to the point that the word is popping up in news stories on everything from subprime mortgages to ovarian cancer research and online advertising.

Most people wisely avoid using the term in casual conversation because the average individual has just a vague idea of what it means. "It's some kind of math that I have no idea how it works," was Tim Lyons' honest response when the J.C. Penney spokesman was asked for his first reaction. He added, "I know the word gets thrown around with Internet-related stuff."

University of Pittsburgh professor Kirk Pruhs, who teaches about algorithms, tries to motivate his students to stay focused on the subject by doing a little name-dropping. He mentions Yahoo!, Amazon, Google, Microsoft, noting the companies "all ask algorithmic questions as part of the interview process." Those questions can be helpful in understanding what the computer scientists mean by thinking abstractly.

Consider this question from www.softwareinterview.com: You have five jars of pills. Each pill weighs 10 grams, except for contaminated pills contained in one jar, where each pill weighs 9 grams. Given a scale, how could you tell which jar had the contaminated pills in just one measurement?

Or this: You've got someone working for you for seven days and a gold bar to pay them. The gold bar is segmented into seven connected pieces. You must give them a piece of gold at the end of every day. If you are only allowed to make two breaks in the gold bar, how do you pay your worker?

The point is to see how the job applicant approaches the problem. An increasing number of companies need people who know their math but also can look at real-world situations and convert them into problems a computer can work on.

North Shore software company SmartOps Corp., founded by a Carnegie Mellon professor, is using algorithms to help businesses such as ConAgra Foods and J.M. Smucker improve supply chain operations.

Squirrel Hill-based search engine operator Vivisimo traces its roots to three CMU scientists with an algorithm and a new way of clustering documents.

American Eagle Outfitters Inc., the fast-growing, youth-oriented retailer, uses such tools to determine what jeans to mark down and when, as well as to forecast how many T-shirts a given store might receive, said Michael Rempell, chief supply chain officer.

"In a nutshell, algorithms help us give our customers what they want, when they want it," he said. "Algorithms are an important piece of how using technology can make a business smarter, faster and more efficient."

Every day, smart people are trying to come up with better algorithms. AltaVista had a good start on the search engine market but a company called Google came along and produced better results with its algorithm for rapidly sorting billions of Web pages, noted Pitt's Dr. Pruhs.

As dominant as Google is, he said, someone else could do better by coming up with an even better algorithm. That's one reason Google keeps setting up shop in places with bright minds, including Oakland. Dr. Pruhs estimates the company hires a hefty portion of the nation's Ph.D.s in algorithm.

Not long ago, Google lured one of the industry's big names -- Udi Manber -- away from Amazon, where he had been the chief algorithms officer, not a title seen every day.

One reason there's still more to do is that some problems are so hard companies and researchers simply use algorithms to get a better solution than they had before. The guy who comes up with an even better recipe will be welcomed.

Both FedEx and MapQuest have to make compromises as they try to lay out the best routes. Lately, eliminating left turns has been seen as a way to make drivers' lives easier, said Dr. Pruhs.

Not everyone who likes computers or math likes the kind of abstract puzzling that attracts the algorithm types. The new center for "computational thinking" that CMU and Microsoft are teaming up to create is meant to encourage those that do.

Though the business world has embraced the use of algorithms, companies must keep in mind that each one is only as good as its design and data, said Dr. Ravi.

The program to optimize use of employees will need a human making choices. Just because an algorithm makes it possible to determine what prices certain shoppers are willing to pay for groceries or cars doesn't mean it would be ethical to charge them different rates, he said.

In one ongoing not-so-academic puzzler that those in the field are keeping an eye on, algorithms are used to encrypt credit card data that consumers and businesses send over the Internet. It would be difficult to break the code using computers available today, although theoretically not impossible.

But a few years ago computer scientist Peter Shor discovered that a quantum computer could probably do it. Now the government is trying to build one before someone else does.