Saturday, October 10, 2009

Introducing: Operations Research

At heart, I am a humanities person; that is why this blog has existed for as long it has. But I do lead another, very different life that consumes most of my waking time. I am an academic in a field called “operations research” – a baffling term for those unaware of its existence. Briefly, operations research can be described as “the science of efficiency”, or “the science of planning well”; the most popular version, though, is the glib-sounding “the science of better”.

Still puzzled? Well, let me try again, this time with an analogy.

We are surrounded by technological marvels. It seems magical – to me at least – that a plane carrying hundreds of passengers and tons of luggage, actually manages to take flight; that there are such things as wireless phones; that there is a large, scattered yet miraculously unified network called the “Internet”. Amazing right? Each of these applications is possible because of engineering, which makes clever use of the underlying science, be it fluid dynamics, signal processing, or fiber optics. Such engineering is not always obvious, but the lay person is aware that there are specialists -- aerospace engineers, computer scientists, electrical engineers to name just a few -- who make these things work.

In the same way, do you wonder how your FedEx package from the Philippines arrived without delay to the small Midwestern town you live in; how the Netflix movie you ordered gets to your address exactly on the day their email claimed; how large airports, such as Heathrow and JFK, manage their flights, schedules, and air traffic? We take these systems for granted, but they work because they are engineered. This type of systems level engineering – the science of allocation and scheduling in the face of uncertainties and the fluctuating dynamics of supply and demand – is called operations research. In business schools it is called management science. Since it is a less tangible kind of engineering, the lay person is generally unaware of it.

You might argue that many systems are rarely well managed. What’s in a science that produces long lines and sapping delays? True, systems may be dysfunctional because of bad planning but this is not unique to operations research. A mechanical problem – arguably caused by the traditional “nuts and bolts” engineer – can stop a flight from taking off as well. In fact, an operationally conscious airline will have a contingency schedule that minimizes the traveler’s disruption in case of a cancellation. Think of all the flight groundings and cancellations that happened on and post 9/11. Have we given close thought to what it took to bring everything back to normal?

Operations research is a mongrel field. Like other engineers, the operations researcher uses mathematical methods, but she also may dabble in statistics, economics, and computer science. She will also need knowledge of the domain she is working in; and importantly, if her domain involves people, she will need to know that people do not behave as rigidly or rationally as her math models assume. This mongrel quality of the field makes it breathtakingly versatile – applications have advanced well beyond the “operations” realm and have entered even areas such as designing beam angles for radiation therapy. The flip side of the coin, however, is that some think of it as an “anything-goes” field with no real identity.

My work is in healthcare operations research. I look at how medical practices can provide timely care while trying to rein in costs. This coincides with the ongoing upheaval in the US healthcare system. Cost and coverage are major issues and they have the power throw askew the balance of supply and demand and influence the quality and timeliness of care. For example, emergency departments in the United States – one setting which I study in my research – experienced a 32% increase in demand over the last decade. The number annual ED visits in the US went up from 90 million in 1996 to 119 million in 2006. This has led to crowded conditions, especially during late afternoons and evenings. Can better staffing, improved coordination of processes alleviate the long wait times of patients? Perhaps -- at least that is what my hypothesis is.

I have also recently discovered – to my great delight – the many ways in which operations research can intersect with the humanities. Let me list a few examples. The resettlement of refugee farmers in India after the partition of the Indian subcontinent in 1947 was a difficult problem in every sense. Nearly a million people had to be allotted new land, and the partition had been extremely violent. Yet it was successfully done, without computers: a classic example of hands-on operations research in which people management and administrative organization are the main skills. The person who led it was Sardar Tarlok Singh of the Indian Civil Service, a graduate of the London School of Economics.

More recently a Markov model (in more plain terms, a probability model) was used to identify syntactic patterns in the as yet undeciphered Indus script of nearly 3000-3500 years ago -- another unconventional application that has nothing to do with “operations”. I am also fascinated by how humanitarian organizations – the UN World Food Program (WFP), Medicines Sans Frontiers – deliver their services in resource constrained settings; understanding why FEMA messed up post-Katrina; and how the dreaded LTTE efficiently coordinated rescue operations post-Tsunami in Sri Lanka.

In short, there’s plenty to learn and explore.

This post comes as I travel to San Diego for the annual meeting of the Institute of Operations Research and Management Science (INFORMS09). Nearly 4000 people will attend the conference; it’s a great way to catch up with friends from graduate school and make new friends. I was also invited to be one of their twelve official bloggers (that gives me an excuse to point to this post there).

This isn’t the first time I’ve mentioned operations research. Here are a few earlier posts: My Adventures During a Queuing Study; Queues and Illegal Immigration; A Visit to an Emergency Room; The Mathematics of Matching Kidneys.