Scientists, Networks, and Cambridge Nights -- But Where are the Females?

In yesterday's Science section of The New York Times there is a very interesting article by John Markoff, Scientist's Online Interviews Draw His Peers Out of Lecture Mode, about an initiative at MIT's Media Lab (which I am a big fan of) called Cambridge Nights. This is a series of video interviews with scientists on their lives and research.

The interviews are conducted by Dr. Hidalgo, who is a physicist, and whose dissertation advisor was Albert-Laszlo Barabasi, of network science fame, who is now at Northeastern University. Barabasi spoke in our UMass Amherst INFORMS Speaker Series and, as always, was provocative, energetic, and very informative. The last time that I saw Professor Barabasi was at the Network Science Conference, held at the Media Lab in Cambridge, MA, May 12-14, 2010, where we presented a paper on our critical needs supply chain network design research.

All of the scientists interviewed in the Cambridge Nights videos, to-date, however, are males, which was disappointing to me. Perhaps, the female President of MIT, Dr. Susan Hockfield, could be included in a future interview or how about Dr. Lisa Randall, the physicist at Harvard? I could easily prepare an appropriate list. Having spent a year as a Science Fellow at the Radcliffe Institute for Advanced Study at Harvard University, I can attest to the female brainpower in Massachusetts.

One of the scientists interviewed (not unexpectedly given the above academic genealogical connection) in Cambridge Nights is Dr. Barabasi, and he speaks about how he became interested in networks and how his work evolved from material sciences to network science. I appreciate his honesty regarding how, initially, he had difficulty publishing his papers. He deserves a lot of credit for bringing deep data analysis to network problems. However, in listening to the video, someone might get the mistaken impression that there was no research on networks until the mid 1990s and that is factually incorrect.

In operations research (and even in economics), the modeling, analysis, and solution of network problems has a long history, dating to the 1940s (and I can even go back further in the case of economics and finance and have done so in quite a few articles that I have published).

In terms of transportation alone the book by Beckmann, McGuire, and Winsten (1956) provided fundamental results in capturing complex behavior of people (drivers) interacting with one another on transportation networks, among other innovations. Also, everyone in operations research is familiar with the book by Ford and Fulkerson, Flows in Networks, published in 1962.

By dissertation advisor at Brown University, Dr. Stella Dafermos, was publishing on networks, starting in 1969, and she helped to inspire me, although she died in 1990, to write quite a few books on network themes, beginning with my first book in 1993.

Some of the classical books on networks that are available for download can be found on the Supernetworks Center website.

Impressions of Network Science at NetSci2010

The NetSci2010 conference, at the Media Lab at MIT, which took place May 12-14, 2010, was a very enjoyable experience for several reasons. I had the opportunity to take one of my doctoral students, Min Yu, to her first professional conference in the US. The paper that I presented, "Supply Chain Network Design for Critical Needs with Outsourcing," was based on our joint work with Dr. Patrick Qiang, and it had been, just this past week, accepted for publication in Papers in Regional Science. It was refreshing to see her excitement and enjoyment of the talks and her enjoyment in meeting various speakers. What really pleased me a lot was that we had, over the past several years, hosted several of the organizers and invited speakers in our Speaker Series in Operations Research / Management Sciences at UMass Amherst. It was terrific to see, once again, Professor Barabasi, Professor Alex "Sandy" Pentland, Professor Asu Ozdaglar, and, albeit, just briefly, Professor David Lazer.

The talks were wide-ranging (from mobile phones, to the diffusion of ideas via Facebook, to supply markets and financial markets and volatility, to honest signals, organizations, and productivity, among others) and there was an appreciation for discussions. Also, there were poster sessions. We even met a few members of the operations research community.

Quite a few of the talks were on data-driven topics, whereas our talk focused on optimization and network design, and redesign with an emphasis on critical needs products, such as vaccines, medicines, and food and water supplies for humanitarian operations in the case of disasters.

We were delighted to come across Professor Rae Zimmerman of NYU, who works on network vulnerability, climate science, and cybersecurity at the conference and got a chance to reflect with her even on regional science. We also spoke with Professor Satish Ukkusuri of Purdue University. Given the contributions to networks from the operations research community going back to the 1940s, it is important that that scientific literature, whose applications have impacted transportation, telecommunications, finance, and even supply chains, both in theory and practice, be represented, in the newer "network science" literature.

Will be Speaking on Supply Chain Network Design for Critical Needs at NetSci2010

The NetSci2010 Conference will be taking place next week at the Media Lab at MIT, which is a gorgeous venue. I will be delivering the paper: Supply Chain Network Design for Critical Needs with Outsourcing, which is joint work with Dr. Patrick Qiang and my doctoral student, Min Yu.

The abstract of the paper is below:

Abstract: In this paper we consider the design of supply chain networks in the case of critical needs as may occur, for example, in disasters, emergencies, pending epidemics, and attacks affecting national security. By "critical needs" we mean products that are essential to the survival of the population, which can include, for example, vaccines, medicine, food, water, etc., depending upon the particular application. "Critical" implies that the demand for the product should be met as nearly as possible since otherwise there may be additional loss of life.

The model that we develop captures a single organization, such as the government or a major health organization or corporation that seeks to "produce" the product at several possible manufacturing plants, have it stored, if need be, and distributed to the demand points. We assume that the organization is aware of the total costs associated with the various operational supply chain network activities, knows the existing capacities of the links, and is interested in identifying the additional capacity outlays, the production amounts, and shipment values so that the demand is satisfied with associated penalties if the demand is not met (as well as penalties with oversupply, which are expected to be lower). In addition, the organization has the option of outsourcing the production/storage/delivery of the critical product at a fixed/negotiated price and with the capacities of those entities being fixed and known. The solution of the model provides the optimal capacity enhancements and volumes of product flows so as to minimize the total cost, which we assume to be a generalized cost, and can include time, subject to the demands being satisfied, as nearly as possible, under demand uncertainty.