When I was a doctoral student way back in the 1980s, getting and using data was a tremendous impediment to finishing a dissertation. Data was precious and very difficult to obtain. Even when received, data often was in an unusable form, involving arcane formatting and coding. We had email, but I can’t recall using it very often. I do recall sending paper letters asking for information, with the resulting weeks before getting a response. Fortunately, I was at a top research university in the United States (Georgia Tech): the data situation elsewhere was undoubtedly worse.
The world has certainly changed. Now, companies and organizations are drowning in data with countless systems generating mega-, giga-, tera-, and even petabytes of data. In 2001, when I first put together a data mining course at Carnegie Mellon, I breathlessly talked about how the books at the U.S. Library of Congress held 20 terabytes of data, then an unimaginable number. I can now buy 20 terabytes of storage for my computer for about $500. Companies like Google, Facebook, Baidu, Twitter and many more take in hundreds of petabytes of data per day.
And, excepting privacy restrictions, this data is not slowed down by national borders. While most of us do not have the bandwidth or computing capability to handle petabytes of data, the kilo- or megabytes of data used by most operational research models are much easier to handle. For any place with a reasonable connection to the internet, data is just a few clicks away.
This has been a tremendous boon to international operational researchers. If you are doing research in integer programming, you have immediate access to MIPLIB, a library of challenging mixed-integer programming instances. You can send instances to NEOS, an online server that can solve a huge range of problems, including linear, mixed-integer, semidefinite, and much, much more. Similar data sources and system exist for a wide range of operational research areas. The internet has been a tremendous force for uniting disparate researchers from around the globe.
But companies around the world are faced with a huge problem: what to do with the data. Whether it be the petabytes of a huge, internet-based company or the kilobytes of a locally run firm, companies need to translate their data into information into knowledge into better decisions. And that that challenge is exactly what operational research is all about. We turn data into decisions. And we do it on a global scale.
Many people recognize company’s needs and I see over and over again attempts to turn data into decisions without understanding that there are a set of tools and skills that we have developed over the past 60 years that do exactly that. We, as a field, need to recognize and embrace the changes in the world. The Age of Analytics should lead to the Age of Operational Research.
My question to you is: what can IFORS do to help individuals and national societies bring on the Age of Operational Research? We bring together people at our conferences, we publish results in our journals, we aid in the education of young people through our support of summer and winter schools, we encourage and support the creation of new national OR societies. What else should we be doing?
I welcome your thoughts and comments at firstname.lastname@example.org. And I hope to see many of you at an upcoming conference, be it EURO in Poznan, INFORMS in the US, or any other conference where our paths cross. And don’t forget to put IFORS 2017 in Quebec City on your calendar: July 17-21, 2017. And Seoul 2020!