Operations Research for the Public Interest Conference
A Conference Exploring the Challenges of Using Operations Research Methods to Solve and Improve Complex Social Problems
This June 2010 conference brought together scientists who study some of the most pressing public problems facing society: Security, energy and environment, and health care. Most of the papers presented were created with collaborative work drawing expertise from governmental entities, non-governmental organizations, and international development organization as well as academics. Scientists at the Stanford Graduate School of Business event shared research findings. Papers from the conference will be considered for a special edition of the journal Operations Research.
Some of the key topics explored were:
How to divide resources between prevention and treatment in fighting the HIV epidemic in Africa
(Research done jointly with the United Nations AIDS Program)
Professor of Management Science and Engineering, School of Engineering, Stanford University
Allocating Resources for HIV Prevention Program Scale-UpSignificant scale-up of global HIV prevention and treatment efforts is needed, and OR tools can help determine the optimal allocation of resources. The heterogeneity of the epidemic means that different intervention packages and scales will be most effective in different settings. Moreover, different regions have different sets of accepted and available interventions and different levels of HIV funding; the marginal effectiveness of interventions may diminish as investments in disease prevention increase; and the HIV epidemic is dynamic and nonlinear. Decision-makers often have little guidance as to which packages of interventions, and at what scale, will yield the best results in their particular settings. Thus, there is a critical need for HIV resource allocation models that can bridge the gap between theory and practice. How to plan the routes of clearing convoys ahead of military convoys in Iraq to prevent damages from explosion
(Research done jointly with U.S. military)
Special lecturer and Professor Emeritus, Graduate School of Business, Columbia University Our focus is on how to make best use of scarce route clearance resources to protect high-value assets, such as convoys, against insurgent improvised explosive device (IED) attacks on highways. This work is motivated by problems that have been experienced by coalition military forces in Iraq and Afghanistan since 2004.
A primary means of mitigating this threat is to clear IEDs from roads by searching for and neutralizing them with dedicated “route clearance teams” or RCTs. Such teams are usually composed of engineering and explosive disposal personnel mounted on a number of specially equipped armored vehicles, equipped with technology designed to detect and neutralize IEDs. Because they are relatively slow moving and there are not enough RCTs to clear every road continuously, they must focus on those highway segments with the highest perceived risk to anticipated coalition force traffic. We aim to take an operations research approach to improving RCT allocation decisions.
How to design efficient recycling programs for high tech devices (Research conducted with the State of Washington)
Nancy J. and Lawrence P. Huang Associate Professor of Operations Management, College of Management, Georgia Institute of Technology, Atlanta
Fair and Efficient Implementation of Collective Extended Producer Responsibility Legislation|
Extended Producer Responsibility (EPR) is a policy tool that holds producers financially responsible for the post-use collection, recycling, and disposal of their products. The goals of EPR are two-fold: To reduce the amount of post-consumer waste that is landfilled, and to improve the recyclability and reduce the toxicity of products by forcing producers to internalize the cost of end-of-life processing. Following its adoption in Europe, EPR is rapidly becoming the preferred policy tool in the U.S. for managing electronic waste. While the collective implementation of EPR aims to achieve cost efficiency by combining volumes and capacities, the European experience demonstrates that the lack of fairness in cost allocation can be a significant barrier to implementation.
Below is a description of other presenters and summaries of research that shaped his or her thinking.
Associate Professor of Business at Darden Graduate School of Business, University of Virginia, Charlottesville
A Newsvendor Model with Pricing for Public Interest Goods
A newsvendor model with pricing considers a firm that determines the order quantity and the price of a good in order to maximize the expected profit over a stochastic demand curve. Such models have been applied in both research and practice, and in numerous situations helped firms better align supply with demand, thus increasing profits. There exists , however, a non-insignificant category of goods for which the above approach of maximizing the firm’s profit would typically be deemed unreasonable. These include safety products (e.g., smoke detectors), energy efficient appliances (e.g., water-saving toilets), health-related products (e.g., vaccines), and other goods that have significant impact on society.
Assistant Professor, Department of Industrial and Systems Engineering, College of Engineering, University of Wisconsin, Madison
Designing a New Breast Cancer Screening Program Considering Adherence
The existing mammography screening guidelines worldwide do not fully capture the dynamics of mammography screening. Specifically, these guidelines consider only age, which is just one of the risk factors associated with breast cancer, when making policy recommendations. Understanding that women at the same age groups do not have uniform breast cancer risks suggests that screening strategies tailored to individual risks may be more beneficial in saving the lives of high-risk women while decreasing unnecessary complications in low-risk women. The existing mammography screening guidelines also do not consider adherence behavior of patients when making policy recommendations. The existing guidelines assume 100% adherence to screening recommendations by patients, an assumption which is not supported by observed screening rates.
Julie Swann: Associate Professor, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology, Atlanta
Two Queue or Not Two Queue: When and How to Integrate HIV Care and Treatment into Outpatient Services in Resource-Limited Settings
Sub-Saharan Africa bears a disproportionate burden of the current global HIV/AIDS epidemic. In 2008, the region accounted for 67 percent of HIV infections and 72 percent of the AIDS-related deaths worldwide. Rapid growth in international donor funding to combat the HIV epidemic has placed an enormous additional strain on already weak public health systems and fueled the debate over vertical versus integrated (or horizontal) health systems and their pros and cons. International donors and their implementing partners have typically favored vertical systems as they enable rapid scale-up of higher quality and more reliable delivery of care in the short term, bypassing potential bottlenecks in public health systems. In the longer term, however, vertical systems can lead to the diversion of human and material resources towards individual diseases, potentially harming overall primary health outcomes for the future. Moreover, disease-specific international funding may not be sustained at its current levels. As a result, there has been a growing need for evidence of feasible integration strategies for HIV services in resource-limited settings. Very little work has been done to evaluate the impact of integration on service delivery systems themselves, specifically focusing on integration of HIV and outpatient care.
Associate Professor, Krannert School of Management, Purdue University, West Lafayette, IN
Building Reliable Air Travel Infrastructure Using Stochastic Models of Airline Networks
Flight delays have been a growing issue and they have reached an all-time high in recent years, with the airlines’ on-time performance at its worst level in 2007 since 1995. A recent report by the Joint Economic Committee of the U.S. Congress has estimated that the total cost to the U.S. economy due to flight delays was as much as $41 billion in 2007. The goal of this paper is to build stochastic models of airline networks and utilize publicly available data to answer the following policy questions: Which network-based, passenger-centric metrics could be used by the FAA to measure on-time performance and schedule robustness? Which are the bottleneck airports in the U.S. air travel infrastructure (i.e., airports that cause most delay propagation)? How would increasing airport capacity at these airports alleviate delay propagation? Which airlines have the least robust schedules? How could these schedules be made more robust?
Orkand Corp. Professor of Management Science holding a joint appointment in Robert H. Smith School of Business, and Institute for Systems Research, University of Maryland, College Park
Social Welfare Justification for Market-Based Approaches for Airport Congestion Management
Over the past four years, the U.S. government has sought to manage congestion at airports in the New York region. These initiatives sought to control congestion and, at the same time, free up capacity to allow access to new and emerging air carriers by auctioning a portion of the airport slots. We investigate other implications of airport congestion management and the alternative mechanisms that might be used to restrict the level of operations at an airport. These lead to an analysis of the tradeoffs between administrative and market mechanisms.
Keynote Speaker - Edward Kaplan
William N. and Marie A. Beach Professor of Management Sciences, Professor of Public Health, and Professor of Engineering, Yale School of Management
This talk considers applications of operations research to intelligence problems in national security and counterterrorism. The phrase “intelligence operations research” can be interpreted in two different ways: as intelligence operations research, meaning studies meant to characterize and improve the operations of intelligence agencies themselves, and as intelligence operations research, meaning the application of operations research methods to specific substantive problems faced by intelligence agencies.
Dick den Hertog
Professor Operations Research, Department of Econometrics and Operations Research, CentER, Tilburg University, Tilburg, The Netherlands
Safe Dike Heights at Minimal Costs, Parts I and II
Protection against flooding is an important issue in the Netherlands since 60 percent of this country is submersible. Each year the Dutch government spends euro1 billion on dikes and dunes. The Dutch Act on the Water Defences gives standards for the maximum flood probability for all dikes in the Netherlands. To satisfy these standards the dikes have to be heightened regularly. Such a heightening is called an update of the dike.
Recently a committee reported that at least 19 percent of all Dutch dikes do not satisfy the standards, and that the standards stated in the above mentioned law were too low, not only because of unforeseen rises in the sea level, but also because economic growth was not taken into account in the cost-benefit analysis. The government therefore asked the authors to develop a new model to determine the right safety standards.
Technical Staff Member, Los Alamos National Laboratory, Los Alamos NM
Strategic Planning for Disaster Recovery with Stochastic Last Mile Distribution
This paper considers the single commodity allocation problem (SCAP) for disaster recovery, a fundamental problem faced by all populated areas. SCAPs are complex stochastic optimization problems that combine resource allocation, warehouse routing, and parallel fleet routing. Moreover, these problems must be solved under tight runtime constraints to be practical in real-world disaster situations. This paper formalizes the specification of SCAPs and introduces a novel multi-stage hybrid-optimization algorithm that utilizes the strengths of mixed integer programming, constraint programming, and large neighborhood search. The algorithm was validated on hurricane disaster scenarios generated by Los Alamos National Laboratory (LANL) using state-of-the-art disaster simulation tools. This work is deployed as part of the National Infrastructure Simulation and Analysis Center (NISAC) and is being used to aid federal organizations such as the Department of Energy and the Department of Homeland Security in planning and responding to disasters.
PhD student in Department of Computer Science, University of Southern California
A Branch and Price Approach to Solving Security Games with Arbitrary Scheduling Constraints
Algorithms to solve security games have seen successful real-world deployment by Los Angeles International Airport police and the Federal Air Marshal Service. These algorithms provide randomized schedules to optimally allocate limited security resources for infrastructure protection. Unfortunately, these algorithms fail to scale-up or to provide a correct solution for massive security games with arbitrary scheduling constraints. This paper provides ASPEN, a branch-and-price algorithm which exploits an innovative compact network flow representation, avoiding a combinatorial explosion of schedule allocations. ASPEN also uses a novel upper-bound generation to speed up the branch and bound. ASPEN is the first method for efficiently solving real-world-sized security games with arbitrary schedules.
Assistant Professor, Decision, Risk, and Operations, Columbia Business School, Columbia University, New York
Structural Estimation of the Chilean Auction for School Meals
The Chilean government provides breakfast and lunch for 2.5 million school children daily. In a developing country where about 14 percent of children under the age of 18 live below the poverty line, many students depend on these free meals as a key source of nutrition. Catering firms that provide this service are awarded contracts through a single round, sealed-bid, first-price combinatorial auction (CA). Meal services are standardized and firms compete in prices. The CA has been used every year since its inception, awarding more than US$3 billion of contracts. Practical experience and academic research have shown that the design of an auction may have an important impact on its outcome -- good designs can result in large savings, while poor designs may lead to large losses for the auctioneer. Typically, the performance of an auction is critically determined by the firms’ cost structure and their strategic behavior. This work analyzes and suggests important improvements to the auction design.
Assistant Professor, Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN
Optimal Rehabilitation Management for Moderate to Severe Traumatic Brain Injury Patients from a Societal Perspective
Traumatic brain injury (TBI) is one of the leading causes of death among Americans under age 45. Although advances in medical technology and treatment have increased TBI survival significantly, little can be done to reverse the initial brain damage caused by trauma and these patients are often left with significant physical and cognitive disabilities. They typically require increased levels of financial and resource commitment during their rehabilitation process, which presents great medical, financial, and social challenges. Wider awareness of the problem and its enormous consequences has led to great interest in developing effective intervention strategies along the TBI rehabilitation process. The post-injury course of TBI patients can be categorized into three phases: treatment, rehabilitation, and survivorship. Rehabilitation requires a continuum of care that can last years and can be costly. For a public funding source, the ultimate goal is to maximize the aggregation of some functional status measure over a cohort of TBI patients it supports. In this proposal, we formulate the problem with an infinite-horizon Markov decision process (MDP) model that includes a set of recurring and absorbing states for each survivor, considers the average of expected aggregate functional status over the patient cohort, and more importantly, includes several budgetary constraints that reflect various consideration of linking individuals in the cohort.
Associate Professor, Department of Mechanical and Industrial Engineering, University of Massachusetts, Amherst
Optimal Climate Change Policy: R&D Investments and Abatement Under Uncertainty
Global climate change is potentially one of the most important environmental, economic, and political problems facing the world today. The rapid warming seen over the past century has provoked fears of possibly catastrophic results from human-induced climate change. Yet, the uncertainty surrounding climate change policy is great. Near-term decisions must be made in the face of, —among others, —uncertainties in technological advance and the ultimate human welfare impacts of increasing concentrations of greenhouse gases. This uncertainty presents a challenge to those of us in the field who would like to provide a rigorous underpinning for climate change policy.
This proposed paper focuses on technological change, in particular on government-funded R&D, and tries to answer the following important public policy question: Given the uncertainty defined by currently available data in technological success and climate change, what are optimal investment policies that maximize social welfare?
Assistant Professor, Industrial and Systems Engineering Department, Pontificia Universidad Católica de Chile, Santiago
Effect of Delays in the Connection-to-the-Grid Time of New Generation Power Plants over Transmission Planning
The start-up times of new power plants are subject to significant uncertainty in deregulated electricity markets and they have an important impact on the planning of the transmission network. In order to measure the effect on transmission planning of postponing the connection time of some generation power plants, we must first develop a model of transmission planning that considers the uncertainty in the instant when new power plants are connected into the electricity grid.
Using models based on Robust Optimization way of dealing with different sources of uncertainty. Robust Optimization seeks solutions for an optimization problem that are insensitive or “immune” to variations in problem data, within certain ranges. The simplest way of using this technique is adopting the worst-case solution. In this article, we propose a model for the electricity network planning inspired in the Robust Optimization approach, which incorporates the uncertainty related with the time in which the new power plants will be connected.
Research assistant at the Department of Business Administration, University of Vienna
Optimizing the Day-Ahead Bidding Strategy of Electricity Storage Using Approximate Dynamic Programming
In wholesale electricity markets, prices and price volatility are primarily driven by time-varying electricity consumption and marginal generation cost that increases in electrical load. A more recent source of price volatility is the large-scale deployment of renewable energies with power generation dependent on changing weather conditions. This raises interest in the economic potential of energy storage as means to improve the flexibility of supply and demand. A power system with a large amount of storage capacity is therefore going to experience a decrease in price volatility with significant external welfare benefits. Operations planning makes storage more efficient and helps fully exploit its economic potential. We examine the bidding process and propose a new algorithm to solve the stochastic-dynamic decision problem underlying storage operation.
Assistant Professor, Systems and Industrial Engineering, University of Arizona, Tucson
Reclaimed Water Network Design under Temporal and Spatial Growth and Demand Uncertainties
Diminishing supplies and population growth are stressing the limited water resources available in many communities. The water, wastewater, and water reuse industries have recognized the need for extending the present water supplies. In many areas, the last remaining untapped water resource is reclaimed or recycled water—treated wastewater that is re-introduced for various purposes. Utilities are under pressure to operate water and wastewater infrastructures in a sustainable manner while providing resilient and robust service. The implication of this approach is that, rather than providing a single water of drinkable quality for all uses, multiple water streams would be distributed to different users in different piping systems. A wide-ranging set of complex problems arise from this new paradigm such as where to locate water and wastewater treatment facilities, how to distribute it and how to plan for population growth. In this paper, we focus on the optimal cost-effective design for the reclaimed water distribution system in Tucson over 20 years.
Assistant Professor, Foster School of Business; University of Washington, Seattle
Influenza Vaccine Supply Chain with Multiple Agencies
Influenza is an acute respiratory illness that spreads rapidly in seasonal epidemics. Annually influenza outbreaks result in 250,000 to 500,000 deaths around the globe. The World Health Organization reports that the costs of health care, lost days of work and education, and social disruption are between $1 million and $6 million per 100,000 inhabitants yearly in industrialized countries. Vaccination is seen as a principal means of preventing influenza, slowing the spread of the disease and even containing a global pandemic. Not only effective, vaccination is also cost-effective and one researcher found that immunization in the elderly saved $117/person in medical costs.
Despite their efficiency and cost-effectiveness, global vaccine allocations to various countries have not been socially optimal. Such sub-optimal vaccine allocations can be attributed to misaligned incentives of decision-makers. The main contribution of this paper is to design contractual agreements between governments to align their incentives and achieve the global optimum solution.
Assistant Professor of Operations Research, Operations Research Department, Naval Postgraduate School, Monterey, Calif.
When a Bit More is Too Much: The Cascading Effect of Insurgency Actions on Popular Attitudes
Several recent military conflicts (Somalia, Iraq, Afghanistan, Southern Lebanon) differ from traditional force-on-force engagements, taking the form of an asymmetrical counterinsurgency (COIN) situation where the civilian population plays a major role. Consequently, legacy warfare and combat models must be updated to reflect this significant shift in conflict paradigm. Almost all current research efforts in COIN modeling concentrate on simulations that heavily rely on data, which is often incomplete or unavailable. This paper focuses on a macroscopic analytic model that relies on general and robust qualitative assumptions, and on relatively small number of key parameters that can be estimated from available data or expert opinions. The objective of the model is to provide transparent cause-and-effect insights about general relations in COIN situations, without imposing unreasonable demands on data. In this paper we focus on the effect of insurgency actions, in particular coercion, on the population's attitude towards the insurgents.
Assistant Professor, Department of Industrial and Systems Engineering, Rensselaer Polytechnic Institute, Troy, NY
Restoring Infrastructure Systems: An Integrated Network Design and Scheduling Problem
We consider the problem of restoring services provided by civil infrastructure systems after an extreme event disrupts them. The public not only hopes for a timely restoration of critical services, such as power, telecommunications, and transportation, but requires them in order for society to recover from the extreme event. Therefore, the managers of the infrastructure systems are faced with demanding choices as they restore these essential public services. |
This research considers a novel integrated network design and scheduling problem that can serve as a decision aid to formulate the restoration efforts in these infrastructure systems. We develop new optimization methods that integrate concepts from the areas of network flows and scheduling. Our methods are tested on realistic data sets representing the infrastructure systems of New Hanover County in North Carolina and lower Manhattan. These results indicate that our methods can be used in real-time restoration planning activities.
Assistant Professor, Department of Industrial and Systems Engineering, State University of New York, Buffalo
Game Theory or Not Game Theory? Hybrid Defensive Resource Allocations
The applicability of game theory to the analysis of terrorism and counter-terrorism has been questioned, with some analysts recommending the use of probabilistic risk analysis or decision analysis instead of game theory, in part on the grounds that it may be unrealistic to assume that terrorists are fully rational or strategic in the game-theoretic sense. Moreover, the question of what to do in the face of an unknown adversary, who may or may not be fully strategic, has not been extensively studied to date. In light of this ongoing controversy, it is significant that we found defensive resource allocations based on game theory to be robust in regards to the possibility of non-strategic attackers. In other words, assuming attackers to be strategic when they are not is conservative, in the sense of incurring less-expected loss than assuming attackers are non-strategic when they are actually strategic. Our results support further development and application of game-theoretical methods, by demonstrating that they can be useful even when the attacker's behavior itself may not be fully strategic or rational.
Charles A. Holloway Professor of Operations, Information, and Technology
Stanford Graduate School of Business
Health Savings Accounts: Consumer Contribution Strategies and Policy Implications
Health Savings Accounts (HSAs) are an alternative to traditional medical insurance plans. They were introduced as part of the 2003 Medicare Modernization Act as a strategy for curbing rising healthcare costs and their status has been preserved in the 2010 Health Insurance Reform bill. In an HSA medical coverage plan, an individual or household makes annual pre-tax contributions to a savings account and tax-free withdrawals to pay for qualified out-of-pocket (OOP) medical expenses.