Jim Orlin, professor of OR at MIT, and coauthor of one of my favorite books (Network Flows: Theory, Algorithms, and Applications with Ravi Ahuja and Tom Magnanti), has just started a blog. It looks like it will be a little more wide-ranging than this one, including education and politics along with operations research. His first OR post is on a great topic: what do you say when someone asks “So what is Operations Research?” Most of us go “well, ummm…, its kinda something with computers” and we lose a great opportunity to educate. So what should we say?
Jim’s preferred definition is “The science of decision making” (I go with something similar: “The science of better decision making”, showing a certain loyalty to the INFORMS “Science of Better” campaign). He follows that up, in the best OR tradition, with an algorithm:
Algorithm for describing operations research to a friend or colleague.
Step 1. Find out a system about which the other person is both interested and knowledgeable. (e.g, sports, entertainment, communication, travel, or anything relating to a person’s job.
Step 2. Develop a plausible scenario based on the system in Step 1; e.g., scheduling sports teams, designing wireless phone systems to provide for the best possible reception, or designing queuing systems at Disneyworld. (I have found that it is very useful to give an example that addresses a problem at the other person’s work that he or she just told you was important.)
Step 3. Explain how operations research can be used to find an excellent solution for the scenario in Step 2 or provide very useful information for the scenario in Step 2.
I like his algorithm, and think it is probably more effective than my version, where Step 1 is “Find a system about which I am knowledgeable and interested…”. My approach tends to lead to a lot of “Oh, look at the time, I must be going!” followed by frantic rushing out of the room. But Jim’s approach does require a bit more thinking on one’s feet: “Oh… so you are interested in the novels of Jane Austen. Well, operations research is, ummm….”
Welcome to the Blog-OR-sphere, Jim!
Jane Austen’s OR application is obviously stable marriage. 🙂 JCB
Perfect! That does go to show that everything is operations research!
I usually use an LP example. A farmer wants to feed his cows as cheaply as possible. He knows the three types of feed available. He knows the nutritional requirements of cows and the vitamins in the feed….
I sometimes preface this by saying that Operations Research is ‘practical math’. Usually this and the above example, that’s about all the people at cocktail parties can stand to feign interest in.
Sometimes I’ll meet an inquisitive person who isn’t just feigning interest and present some n-p complete problem, like the traveling salesman… I usually use ‘find the shortest garbage collection route’.
Farmers & garbage — that’s stuff people usually know about or can imagine. Having ‘back-pocket’ answers like this with common domain problems that are quick to deploy are effective at conveying the subject matter & complexity. That’s typically highly effective at moving on to the next topic of the evening, like “have you seen the latest Battlestar Galactica?” or “who’s going to the superbowl?” YMMV.
Maybe it’s different at the university, but trying to guess or interpret what people are interested in is a mixed bag. Sometimes people don’t want their problems solved at a cocktail party. And they don’t want to think… Answering their question about what is O.R. by immediately questioning their background /interests may be an initial obstacle to discerning their interests/background and prevent you from openings later on where you could inject a little intro of ‘Science’ into their decision problems.
Finding people who not only want to listen to formulation but also the various approaches to a solution are rare indeed. They can often turn into treasured lunch buddies.
All, IMHO, of course.
How great is this blog post. I too am asked that very question of what is Operations Research. I typically say it is “using mathematics and statistics to help solve business decisions”. I like the “science of better decision making” a little bit more. I’ll have to use that.
Can you feed more examples about LP so I can understand the concept – and its applications more?
You may want to leave more examples here, or try to go to my blog. (though it’s not an OR blog).
Thanks.