Grand Challenges

Report outlines operations research’s potential role as catalyst for addressing engineering technological challenges.

By Suvrajeet Sen, et al.

The success of operations research (O.R.) has been built on the ability of the field to transcend disciplinary boundaries in making fundamental research contributions using O.R. modeling approaches and methodologies. The Grand Challenges of the U.S. National Academy of Engineering (NAE) present an opportunity for the O.R. community to play the role of a catalyst, utilizing O.R. to facilitate some of the pressing technological challenges facing humanity today. This is the first of a series of articles that will explore potential avenues for future research in the key areas outlined in the NAE Grand Challenges. What follows is a condensed and edited version of the first portion of a report submitted to the National Science Foundation (NSF) in May 2014 (Sen, et al. 2014).

Background

A panel of thought-leaders convened by the NAE (and facilitated by NSF) unveiled its vision of the Engineering Grand Challenges in 2008. Over the intervening years, this report has invited (and received) feedback from international leaders and professional organizations, including the Institute for Operations Research and the Management Sciences (INFORMS). As part of the INFORMS input, Barnhart et al. (2008) prepared a report on the role that the O.R. community was likely to play within the context of the challenges. An abbreviated version of that report appeared as the President’s Desk column in OR/MS Today (April 2008).

As predicted, the O.R. community has been active in many of the thematic areas of the NAE Grand Challenges via publications on the highlighted research areas in our flagship journals, major conferences on topics at the intersection (e.g., the joint INFORMS-Medical Decision-Making Conference in Phoenix 2012) and several thematic conferences on the smart grid, homeland and cyber security, and others. Because O.R. brings together a combination of tools from computing, mathematics and economic sciences, such an effort is likely to unleash a vast array of new approaches onto the engineering grand challenges of today.

One of the distinctions between O.R. and other mathematical sciences is that many of our premier journals are not only devoted to novel O.R. methodology, but also to real-scale applications. This breadth prompted George Dantzig to label O.R. as a “can-do” discipline, which is the likely reason for major O.R. conferences that cut across both theory and applications, covering mathematical, algorithmic, organizational and practical implementations of O.R. This unique style of discovery may also be traced to the fact that, as a field, O.R. is applied daily to an amorphous body of applications covering a gamut of domains, including the Grand Challenges. For these reasons, companies that are interested in research at the intersection of business analytics, computing and mathematical sciences routinely appoint doctorates from the field of O.R. to lead large research labs (e.g., IBM’s T.J. Watson Research Center).

Major users of O.R., such as the Department of Defense, are clamoring for advanced O.R. tools for trade-off studies (Defense Science Board Report, 2011, especially Chapter 3). Other agencies have also found great value in O.R. tools, as exemplified in the following passage from the 2010 President’s Council of Advisors on Science and Technology (PCAST) report (Holdren, et al., 2010, pp. 71):

“Progress in Algorithms Beats Moore’s Law”

“Everyone knows Moore’s Law – a prediction made in 1965 by Intel co-founder Gordon Moore that the density of transistors in integrated circuits would continue to double every 1 to 2 years.

Fewer people appreciate the extraordinary innovation that is needed to translate increased transistor density into improved system performance. This effort requires new approaches to integrated circuit design, and new supporting design tools, that allow the design of integrated circuits with hundreds of millions or even billions of transistors, compared to the tens of thousands that were the norm 30 years ago. It requires new processor architectures that take advantage of these transistors, and new system architectures that take advantage of these processors. It requires new approaches for the system software, programming languages, and applications that run on top of this hardware. All of this is the work of computer scientists and computer engineers.

Even more remarkable – and even less widely understood – is that in many areas, performance gains due to improvements in algorithms have vastly exceeded even the dramatic performance gains due to increased processor speed.

Making sustainability energy economical is a promising area for O.R. as a catalyst.

The algorithms that we use today for speech recognition, for natural language translation, for chess playing, for logistics planning, have evolved remarkably in the past decade. It’s difficult to quantify the improvement, though, because it is as much in the realm of quality as of execution time.

In the field of numerical algorithms, however, the improvement can be quantified. Here is just one example, provided by Professor Martin Grötschel of Konrad-Zuse-Zentrum für Informationstechnik Berlin. Grötschel, an expert in optimization, observes that a benchmark production planning model solved using linear programming would have taken 82 years to solve in 1988, using the computers and the linear programming algorithms of the day. Fifteen years later – in 2003 – this same model could be solved in roughly 1 minute, an improvement by a factor of roughly 43 million. Of this, a factor of roughly 1,000 was due to increased processor speed, whereas a factor of roughly 43,000 was due to improvements in algorithms! Grötschel also cites an algorithmic improvement of roughly 30,000 for mixed integer programming between 1991 and 2008.”

The benchmark that Grötschel used was drawn from Bixby (2002), published in the INFORMS flagship journal Operations Research. These advances are more than just new computational benchmarks; they have enormous economic impact as well. For instance in the electrical power sector, O.R. models have transformed market operations with advanced software for scheduling, economic dispatch, power flow and many other functions. A recent Federal Energy Regulatory Commission (FERC) report (FERC 2011) attributes savings in the range of $100 million annually for generator scheduling in just one of the markets (PJM, Pennsylvania, New Jersey, Maryland).

These vignettes suggest that we are witnessing a convergence of systems and computational thinking, which has made O.R. a central enabling technology for a variety of applications. Nevertheless, the level of penetration of O.R. in some important domains of science and engineering research has been somewhat limited. For far too long, researchers in science and engineering have failed to recognize the need for scalable models and algorithms. With the increasing need for guided scientific exploration, O.R. approaches for scalable models and algorithms are becoming indispensable. Given the magnitude of problems like climate change, O.R. modeling should be playing a bigger role in understanding the impact of human choices on the future of the planet.

Among the advantages that the O.R. community can bring are tools that integrate data and decisions. This interplay is the key distinction between O.R. models and descriptive statistics. Moreover, the former also facilitates risk modeling in a resource-constrained setting. The time has come to engage both domain experts as well as O.R. experts, so that policies and decisions become an integral part of analysis, not an afterthought. Such collaboration has the potential to discover strategies to reverse the ominous climate-change trends that have been observed over the past two decades.

The NAE Grand Challenges were broadly classified into four categories: Sustainability, Security, Human Health and Joy of Living. The report by Sen, et al. (2014) described potential areas of intersection with and roles that O.R. can play in addressing the challenges in these domains. This article will summarize the four areas, and subsequent articles will address each in more detail. Please see the report for a fuller set of references, which have been abbreviated here.

O.R. for Sustainability

Visualizing the growth of operations research.

Foremost among the challenges are those that must be met to ensure the future itself. The Earth is a planet of finite resources, and its growing population currently consumes them at a rate that cannot be sustained. Utilizing resources like fusion, wind and solar power, preserving the integrity of our environment and providing access to potable water are the first few steps to securing an environmentally sound and energy-efficient future for all of mankind. Analyzing these challenges includes using data and models to choose among alternative strategic decisions, forecasting the effect of decisions on the future, and quantifying the uncertainty associated with this analysis.

While O.R. methods certainly support many individual features (e.g., nonlinear, dynamic, stochastic and discrete), combinations of these features are often necessary in many of the research questions that arise under the “Sustainability” banner. Combining these features will require significant extensions of the O.R. methodology available today. The following areas in sustainability were found to be particularly promising for O.R. as a catalyst:

  • making solar energy economical,
  • integrating storage with solar cells,
  • providing energy from fusion,
  • managing the power grid,
  • geological carbon sequestration,
  • managing water availability and quality, and
  • other social sciences/economics considerations.

O.R. for Security

The ongoing integration of the many infrastructure systems in the modern world – transportation, energy, water, communications, finance – has made these systems more vulnerable to both unintentional hazards and intentional threats. O.R. researchers and practitioners build operational models of such systems precisely because we know that the system’s performance can depend, often in surprising and subtle ways, on how various components and subsystems interact. As our interconnected systems grow in complexity, having a trusted operational model is even more essential for assessing system vulnerabilities and in turn addressing the challenge of how to secure that system.

O.R. is ideally positioned to address the following challenges: formulating operational models of appropriate fidelity; understanding vulnerabilities using models for chance hazards and malicious attacks; and allocating scarce resources to best secure systems. The following areas in security were found to be particularly promising for O.R. as a catalyst:

  • restoring/modernizing critical urban infrastructure systems,
  • preventing nuclear terrorist attacks,
  • enabling/enhancing cybersecurity, and
  • enhancing aviation safety, both on the ground and in the air.

O.R. for Human Health

Healthcare expenditures continue to rise both in absolute terms and as a percentage of U.S. spending. Both improved treatment and reduced expenditures are possible with proper analysis. In many cases, this requires studying the parts of the system as a whole. In other cases, individual procedures, treatments and medicines can be improved through analytical insights aided by data. In any case, the discipline of O.R. brings ideal tools to aid experts in traditional domains of the nation’s healthcare system. The following areas were found to be particularly promising:

  • advanced health informatics,
  • engineering better medicines,
  • reverse-engineering the brain, and
  • improving quality while reducing costs.

O.R. for Joy of Living

Of the four broad categories of NAE challenges, this area is the most nebulous for scientists and engineers, and in the report (Sen, et al., 2014), “Joy of Living” was interpreted as consisting of three core problem areas: 1) advanced personalized learning, 2) enhancing virtual reality, and 3) engineering the tools of scientific discovery. However, one recognizes that “Joy of Living” encompasses a much broader class of challenges dealing with improving the quality of life on a daily basis. For example, reducing traffic congestion in urban areas, improving response times of first-responders, designing smart, energy-efficient homes and others raise many interesting O.R. questions.

One such example is an application related to predicting movie recommendations associated with the so-called “Netflix Prize” problem. This problem is concerned with a matrix whose columns represent “user names,” and the rows represent “movie names.” The problem of predicting which movies should be recommended to a user can be formalized as a “matrix completion” problem of inferring some entries, based on partial data about movie likes/dislikes of some users. Among the more widely cited approaches to this problem is a method by Candes and Tao (2005) where the term “Dantzig-selector” was coined for a convex (linear) programming formulation of this problem.

Other machine learning approaches routinely draw upon optimization as a core technology for inference. Other “joys of life,” such as sports, have also seen many applications of analytics; in addition to the well-publicized baseball movie “Moneyball,” there is Major League Baseball scheduling, which is done routinely using O.R. models. In this sense, O.R. casts such a wide net in the “Joy of Living” area.

Conclusions

O.R. has a long history of bringing analytics to real-world problems. The report by Sen, et al. (2014) is intended to accelerate collaboration between science/engineering on the one hand, and the computational, mathematical and economic sciences on the other. Trial-and-error experimentation is giving way to greater focus on modeling, optimization and simulation as methodologies for seeking new discoveries in energy, materials, nano-technology, medicine and, of course, manufacturing systems and supply chains. In addition, the analytics boom propelled by O.R. is taking business enterprises to new levels of competitiveness. This combination of innovation and competitiveness bodes well for U.S. industry, and can only be sustained by greater cross-fertilization between O.R. and the more traditional areas of science and engineering.

The growth of O.R. may be visualized in the form of the graphic depicted in Figure 1. The innermost (light brown) circle holds the early disciplines that motivated O.R. These include computing/communications, finance, military, transportation and others. As a result of its support of these “traditional applications,” the field of O.R. has created its own “ring” of concepts that integrate several fundamental pillars of O.R. knowledge. We expect that the next phase of O.R. growth will result from greater exchanges with domains associated with the Grand Challenges (which appear on the outer edges). Over time, these challenges will add new dimensions to O.R. that will then be represented by other new “rings” of O.R. concepts. This approach, which we refer to as multi-disciplinary operations research and engineering (MORE), will not only lead to new science and engineering knowledge, but to new transformative technologies and new O.R. as well.

Suvrajeet Sen is a professor in the Daniel J. Epstein Department of Industrial and Systems Engineering, Viterbi School of Engineering, University of Southern California. He led the team that produced the report, “O.R. as a Catalyst for Engineering Grand Challenges,” upon which this article was based. See the box below for the other members of the report team.

Contributors to the Report

“O.R. as a Catalyst for Engineering Grand Challenges,” a report to the National Science Foundation, was compiled by a team of contributors led by Suvrajeet Sen of the University of Southern California. Other contributors, all of them prominent members of INFORMS and the worldwide O.R. community, included:
  • Cynthia Barnhart, Massachusetts Institute of Technology
  • John R. Birge, University of Chicago
  • E. Andrew Boyd, PROS
  • Michael C. Fu, University of Maryland
  • Dorit S. Hochbaum, University of California, Berkeley
  • David P. Morton, Northwestern University
  • George L. Nemhauser, Georgia Institute of Technology
  • Barry L. Nelson, Northwestern University
  • Warren B. Powell, Princeton University
  • Christine A. Shoemaker, Cornell University
  • David D. Yao, Columbia University
  • Stefanos A. Zenios, Stanford University

Notes & References

  1. Barnhart, C., Daskin, M.S., Dietrich, B., Kaplan, E., and Larson, R., 2008, “Grand challenges in engineering,” INFORMS white paper, submitted as response to NAE Grand Challenges.
  2. Bixby, R.E., 2002, “Solving real-world linear programs: A decade and more of progress,” Operations Research, Vol. 50, pp. 3-15.
  3. Candes, E.J. and Tao, T., 2005, “Decoding by linear programming,” IEEE Transactions on Information Theory, Vol. 51, pp. 4203-4215.
  4. Defense Science Board Report, 2011, “Enhancing Adaptability of U.S. Military Forces,” p.31.
  5. FERC Staff Report, 2011, “Recent ISO software enhancements and future software and modeling plans.”
  6. Holdren, J.P., Lander, E. and Varmus, H., 2010, “Report to the President and Congress – Designing a Digital Future: Federally Funded Research and Development in Networking and Information Technology,” President’s Council of Advisors on Science and Technology.
  7. Sen, S., Barnhart, C., Birge, J.R., Boyd, E.A., Fu, M.C., Hochbaum, D.S., Morton, D.P., Nemhauser, G.L., Nelson, B.L., Powell, W.B., Shoemaker, C.A., Yao, D.D. and Zenios, S.A., 2014, “O.R. as a Catalyst for Engineering Grand Challenges,” Report to the National Science Foundation, Arlington, Va.; available for download at http://connect.informs.org/communities/community-home/librarydocuments/viewdocument/?DocumentKey=d7e454e3-1872-4826-900b-7871063a5980.