Applied Research

Have you ever read a research paper that began something like the following?:

“Consider the situation were there is a single customer, and exactly two producers of widgets. Demand for widgets follows a poisson distribution with parameter λ. Further, on the first day of each month, each producer sets their price for the coming month without knowing the price of the other producer. Producer 1 has a maximum production capacity of M1, and producer 2 has a maximum production capacity of M2. To meet monthly requirements, the customer will buy as many widgets as possible from the low cost producer. We show that an optimal…”

Regarding papers such as this, Russell Ackoff, in a 1979 paper entitled ‘The Future of Operational  Research is Past’ wrote:

“…engaging in impure research couched in the language, but not the reality, of the real world.”

And, more recently, ManMohan Sodhi and Christopher Tang, in a 2008 paper, entitled ‘The OR/MS Ecosystem: Strengths, Weaknesses, Opportunities, and Threats’ wrote:

“…OR/MS research is retreating from real-world applications.” and

“Without testing in the “real world,” there is no correcting force to prevent OR/MS from becoming “too mathematical”. “

It seems that it has been recognized for some time, that much of the OR/MS research that is conducted, has become disconnected from real-world problems. In the view of Sodhi and Tang, this leads to “excessive self-referentiality”, and research that loses relevance to practitioners and end users. To the extent that this is true, such research will not increase the interest in, or the number of users of, OR/MS.

A Simple Solution

There is however, a simple way for OR/MS journals to encourage more relevant research. Currently, the journal ‘Operations Research’ has two main types of papers: applications and theory papers. Applications are solutions that have actually been deployed by an organization, while theory papers are not required to directly relate to a real-world problem. The overwhelming majority of papers published are theory papers. What is needed, is a third type of research paper: applied research. These papers will not describe deployed applications, but unlike theory papers, they will be required to describe and solve a real-world problem.

The determination as to whether the paper solves a real-world problem will be made during an expanded review process: judgements will be made as to whether the stated problem is an actual real-world problem, and whether the proposed model could actually solve that problem.

The new applied research category can be implemented in a completely positive way: a journal can allocate additional pages to this category, while continuing to publish the same number of pages in the theory paper and application categories. Applied research will become a new option available to prospective authors.

Everybody Wins

As authors begin to reconnect with real-world problems, and experience the benefits of publishing applied research papers — more consulting opportunities and greater recognition of their research — this type of research will become increasingly popular. Moreover, the benefits will not be limited to the authors of applied research papers: the departments in which they reside, operations research practitioners, government and commercial organizations, and the OR/MS profession as a whole, will benefit from increased awareness of the value of operations research solutions.

2 thoughts on “Applied Research

  1. Steve Nahmias

    Bob: I enjoy your frequent comments posted here. I’ve been an academic for 43 years and folks have been talking about this issue a lot longer than that. We can list a lot of reasons why OR has evolved this way, but the fact is that professors have 6 years to develop a publication record and are either given tenure or fired based on what they’ve been able to publish. With turnaround times in our major journals measured in years, young faculty are really under the gun. The fastest road to a publication is to generalize existing theory. Real problems are messy. Data is hard to come by, and is unreliable in many cases. Part of the reason the system has evolved this way is that it works. We need these professors to fill thousands of positions to education millions of students. And finally, some of this useless theory can turn out to be useful down the road.

    Just playing devil’s advocate.

  2. Philbert Morris

    Agree with Steve. I tell people that academics break up the world into Chemistry and Physics etc, and then pretend that these divisions are real When you go out there to solve a problem, the person at the other end does not care WHAT you are using to solve it. If you have to use German to solve it. so be it.
    The academic, on the other hand, has a tool, and his interest is in using it.
    So often you solve an important problem, but what you use has no “publishable” content.
    The only courses that will remain “lifelike” will be the ones that teach the art and science of modelling and problem-solving.

Comments are closed.