Operations Research, Decision Making

Decision making, operations research, transportation science, supply chain and revenue management, network reliability and optimization

ISR is a recognized leader in decision making and operations research. Our faculty and students established a model-based systems engineering approach for integrated product process design, including object-oriented models of system behavior and structure, and optimization-based tradeoff analysis. We also developed formal model checking methodology for validation, verification and safety of hybrid biological and automotive systems. We have developed models for decision making, including for mass vaccination clinics and other public health needs. Our systematic approach for computer-aided manufacturability analysis of machined parts brought the problem of existence of alternative interpretations to the attention of the feature recognition community. We developed efficient algorithms and software, including neural network models, for input-output behavior and model predictive control of chemical processes—making semiconductor wafer fabrication more efficient. Perhaps most significantly, for more than 20 years, as part of the NEXTOR consortium, ISR researchers have conducted operations research for the Federal Aviation Administration in air traffic management and control, aviation economics and policy, and performance evaluation and metrics.

Recent news

Visit the NEXTOR II website

Recent publications

2021

Dynamic Slot Exchange Mechanisms in Air Traffic Management

YMichael Ball, Thomas Vossen

The researchers use the slot exchange concept as a starting point to present and analyze two advanced exchange mechanisms. They replace the current one-for-one exchange mechanism, i.e. compression, with a two-for-two exchange mechanism. Subsequently, they add the possibility of monetary side-payments to the exchange. Implementation of the the two-for-two exchange mechanism requires the solution of a mediator problem, which must determine the set of two-for-two offers to accept. The incorporation of monetary side payments represents a fundamental shift from barter to a true marketplace. Supporting such a marketplace requires the development of an appropriate auction mechanism. In both cases, the effectiveness of the exchange mechanisms should be measured by comparing the efficiency of the allocations the mechanisms can achieve with a globally optimal allocation. Each of these topics represents an interesting research challenge, which the authors address.

TRISTAN V: The Fifth Triennial Symposium on Transportation Analysis

2020

Incorporating User Preferences in Time-Based Flow Management Operations

Yeming Hao, Sergio Torres, David Lovell, Michael Ball

Traffic flow systems that balance demand versus capacity at busy airports assign Controlled Times of Arrival (CTAs) to incoming flights. This paper evaluates a strategy to assign these CTAs based on user-provided priority lists. The user-provided priority is used to drive a slot swapping algorithm that looks for opportunities to rearrange the order of flights in the CTA queue in a way that decreases delay cost. The authors quantify potential savings by comparing the queue after swapping with the default first-come-first-served rule.

2020 AIAA/IEEE 39th Digital Avionics Systems Conference

An Arrival Scheduling Model for Incorporating Collaborative Decision-Making Concepts into Time-Based Flow Management

Yeming Hao, David Lovell, Michael Ball, Sergio Torres, Gaurav Nagle

This paper proposes a flight scheduling scheme–2-opt-swap, which assigns controlled times of arrival (CTAs) for flights reaching the Freeze Horizon and allows certain slot swapping between different flights with the goal of reducing total controlled arrival delay cost over all carriers. The allowable swaps are predicated on models of carrier preferences following a Collaborative Decision-Making paradigm. Monte Carlo simulations were designed to prove the benefits of this new CTA scheduling scheme, compared to a baseline model of first-come-first-served discipline, which is currently used in Time-Based Flow Management.

2020 International Conference on Research in Air Transportation (ICRAT)

2021

Voice Interface Technology Adoption by Patients With HeartFailure:Pilot Comparison Study

Lida Anna Apergi, Margret Bjarnadottir, John Baras, Bruce Golden, Kelley M Anderson, Jiling Chou, Nawar Shara

This is a study of the engagement of patients with heart failure with voice interface technology. In particular, the authors investigate which patient characteristics are linked to increased technology use.

JMIR mHealth and uHealth

 

2020

An optimization model for multi-appointment scheduling in an outpatient cardiology setting

Lida Apergi, Bruce Golden, John Baras, Kenneth Wood

This research tackles the problem of outpatient scheduling in the cardiology department of a large medical center, where patients have to go through a number of diagnostic tests and treatments before they are able to complete a final interventional procedure or surgery. The authors develop an integer programming (IP) formulation to ensure that the outpatients will go through the necessary procedures on time, that they will have enough time to recover after each step, and that their availability will be taken into account.

Operations Research for Health Care

2020

Iterative Pre-Conditioning to Expedite the Gradient-Descent Method

Kushal Chakrabarti, Nirupam Gupta, Nikhil Chopra

The authors propose an iterative pre-conditioning method that significantly reduces the impact of the conditioning of the minimization problem on the convergence rate of the traditional gradient-descent algorithm.

arXiv.org

Decentralized Multi-subsystem Co-design Optimization Using Direct Collocation and Decomposition-based Methods

Tianchen Liu, Shapour Azarm, Nikhil Chopra

Two decentralized (multi-level and bi-level) approaches are formulated to solve multi-subsystem co-design problems, which are based on the direct collocation and decomposition-based optimization methods.

Journal of Mechanical Design

2021

Variance Reduction for Generalized Likelihood Ratio Method in Quantile Sensitivity Estimation

Yijie Peng, Michael Fu, Jiaqiao Hu, Pierre l’Ecuyer, Bruno Tuffin

The authors apply the generalized likelihood ratio (GLR) methods in Peng et al. (2018) and Peng et al. (2021) to estimate quantile sensitivities. Conditional Monte Carlo and randomized quasi-Monte Carlo methods are used to reduce the variance of the GLR estimators. The proposed methods are applied to a toy example and a stochastic activity network example. Numerical results show that the variance reduction is significant.

INRIA's HAL multi-disciplinary open access archive

Adaptive Importance Sampling for Efficient Stochastic Root Finding and Quantile Estimation

Shengyi He, Guangxin Jiang, Henry Lam, Michael Fu

In solving simulation-based stochastic root-finding or optimization problems that involve rare events, such as in extreme quantile estimation, running crude Monte Carlo can be prohibitively inefficient. To address this issue, importance sampling can be employed to drive down the sampling error to a desirable level. However, selecting a good importance sampler requires knowledge of the solution to the problem at hand, which is the goal to begin with and thus forms a circular challenge. We investigate the use of adaptive importance sampling to untie this circularity.

arXiv.org

2020

Predictive Modeling for Epidemic Outbreaks

Jian Chen, Michael Fu, Wenhong Zhang, Junhua Zheng

Describes a new discrete-time Markov chain predictive model for the COVID-19 pandemic that provides a way to estimate parameters from available data and is computationally tractable both in terms of parameter estimation and in model output analysis. This model has been adopted by the first Shanghai assistance medical team in Wuhan’s Jinyintan Hospital, where the forecasts have been used for preparing and allocating medical staff, ICU beds, ventilators, and other critical care medical resources and for supporting medical management decisions.

Asia-Pacific Journal of Operational Research

2021

Locally Optimizable Joint Embedding Framework to Design Nitrogen-rich Molecules that are Similar but Improved

Sangeeth Balakrishnan, Francis VanGessel, Zois Boukouvalas, Brian Barnes, Mark Fuge, Peter Chung

Deep learning has shown great potential for generating molecules with desired properties. But the cost and time required to obtain relevant property data have limited study to only a few classes of materials for which extensive data have already been collected. The authors develop a deep learning method that combines a generative model with a property prediction model to fuse small data of one class of molecules with larger data in another class.

Molecular Informatics

How should we measure creativity in engineering design? A Comparison of social science and engineering approaches

Scarlett R. Miller, Samuel T. Hunter, Elizabeth Starkey, Sharath Kumar Ramachandran, Faez Ahmed, Mark Fuge

Design researchers have long sought to understand the mechanisms that support creative idea development. However, one of the key challenges faced by the design community is how to effectively measure the nebulous construct of creativity. The social science and engineering communities have adopted two vastly different approaches to solving this problem, both of which have been deployed throughout engineering design research. This paper compares and contrasts these two approaches using design ratings of nearly 1000 engineering design ideas. While these two methods provide similar ratings of idea quality, there was a statistically significant negative relationship between these methods for ratings of idea novelty. The results also show discrepancies in the reliability and consistency of global ratings of creativity and provide guidance for the deployment of idea ratings in engineering design research and evidence.

ASME Journal of Mechanical Design

2019

Using Semantic Fluency Models Improves Network Reconstruction Engineering Knowledge

Thurston Sexton, Mark Fuge

The paper directly models a cognitive process by which technicians may record work orders, recovering implied engineering knowledge about system structure by processing written records.

ASME 2019 International Design Engineering Technical Conference/Computers and Information in Engineering Conference

Checking the automated construction of finite element simulations from Dirichlet boundary conditions

Keven Chiu, Mark Fuge

From engineering analysis and topology optimization to generative design and machine learning, many modern computational design approaches require either large amounts of data or a method to generate that data. This paper addresses key issues with automatically generating such data through automating the construction of Finite Element Method (FEM) simulations from Dirichlet boundary conditions.

ASME 2019 International Design Engineering Technical Conference/Computers and Information in Engineering Conference

2020

Lookahead and Hybrid Sample Allocation Procedures for Multiple Attribute Selection Decisions

Jeffrey W. Herrmann, Kunal Mehta

Attributes provide critical information about the alternatives that a decision-maker is considering. When their magnitudes are uncertain, the decision-maker may be unsure about which alternative is truly the best, so measuring the attributes may help the decision-maker make a better decision. This paper considers settings in which each measurement yields one sample of one attribute for one alternative.

arXiv.org

Modeling for emergency public health clinics

Jeffrey W. Herrmann and multiple colleagues

From 2005–2013, Professor Jeffrey W. Herrmann (ME/ISR) worked on public health research with the U.S. Centers for Disease Control and Prevention (CDC); the Montgomery County, Maryland, Advanced Practice Center for Public Health Emergency Preparedness and Response; and the National Association of County and City Health Officials (NACCHO).

Specifically, Herrmann and his colleagues created mathematical and simulation models of mass dispensing and vaccination clinics (also known as points of dispensing or PODs). They also developed decision support tools to help emergency preparedness planners design clinics that have enough capacity to serve residents quickly while avoiding unnecessary congestion.

A poor clinic design has insufficient capacity and long lines of patients waiting for vaccinations or other services such as testing and triage. More patients require more space as they wait to receive treatment. If too many patients are in the clinic, they cause congestion, crowding, confusion, and the potential spread of disease. Herrmann's models and support tools help planners design clinics that have better patient flow and are able to process people more quickly and efficiently.

This research originally was developed with mass dispensing and vaccination clinics in mind. However, the principles behind the models and support tools are useful for those setting up emergency clinics to combat the COVID-19 pandemic. They are offered free of charge.

2020

Incorporating User Preferences in Time-Based Flow Management Operations

Yeming Hao, Sergio Torres, David Lovell, Michael Ball

Traffic flow systems that balance demand versus capacity at busy airports assign Controlled Times of Arrival (CTAs) to incoming flights. This paper evaluates a strategy to assign these CTAs based on user-provided priority lists. The user-provided priority is used to drive a slot swapping algorithm that looks for opportunities to rearrange the order of flights in the CTA queue in a way that decreases delay cost. The authors quantify potential savings by comparing the queue after swapping with the default first-come-first-served rule.

2020 AIAA/IEEE 39th Digital Avionics Systems Conference

An Arrival Scheduling Model for Incorporating Collaborative Decision-Making Concepts into Time-Based Flow Management

Yeming Hao, David Lovell, Michael Ball, Sergio Torres, Gaurav Nagle

This paper proposes a flight scheduling scheme–2-opt-swap, which assigns controlled times of arrival (CTAs) for flights reaching the Freeze Horizon and allows certain slot swapping between different flights with the goal of reducing total controlled arrival delay cost over all carriers. The allowable swaps are predicated on models of carrier preferences following a Collaborative Decision-Making paradigm. Monte Carlo simulations were designed to prove the benefits of this new CTA scheduling scheme, compared to a baseline model of first-come-first-served discipline, which is currently used in Time-Based Flow Management.

2020 International Conference on Research in Air Transportation (ICRAT)

2020

Fair Liver Transplant Allocation: A Scalable Optimization Model

Shubham Akshat, S. Raghu Raghavan, Sommer Gentry

Presents a new nonlinear integer programming model for U.S. liver transplant allocation a model that allocates donor livers to maximize minimum supply/demand ratios across all transplant centers. In simulations, the model reduces disparities and saves lives.

Raghavan preprint website

2019

Least-Cost Influence Maximization on Social Networks

Dilek Günneç, S. Raghavan, Rui Zhang

This paper focuses on the NP-hard “least-cost influence problem (LCIP)”: an influence-maximization problem where the goal is to find the least-expensive way of maximizing influence over a social network.

INFORMS Journal on Computing, Nov. 26, 2019

2021

Policy Optimization in Bayesian Network Hybrid Models of Biomanufacturing Processes

Hua Zheng, Wei Xie, Ilya Ryzhov, Dongming Xie

Biopharmaceutical manufacturing is a rapidly growing industry with impact in virtually all branches of medicine. Biomanufacturing processes require close monitoring and control, in the presence of complex bioprocess dynamics with many interdependent factors, as well as extremely limited data due to the high cost and long duration of experiments. The authors develop a novel model-based reinforcement learning framework that can achieve human-level control in low-data environments.

arXiv.org

 

2020

Personalized Multimorbidity Management for Patients with Type 2 Diabetes Using Reinforcement Learning of Electronic Health Records

Hua Zheng, Ilya Ryzhov, Wei Xie, Judy Zhong

Comorbid chronic conditions are common among people with type 2 diabetes. The authors developed an Artificial Intelligence algorithm, based on Reinforcement Learning, for personalized diabetes and multimorbidity management with strong potential to improve health outcomes relative to current clinical practice.

Operations Research

Consistency Analysis of Sequential Learning Under Approximate Bayesian Inference

Ye Chen, Ilya Ryzhov

Introduces a new consistency theory that interprets approximate Bayesian inference as a form of stochastic approximation (SA) with an additional “bias” term. Derives the first consistency proofs for a suite of approximate Bayesian models.

Operations Research

2020

An optimization model for multi-appointment scheduling in an outpatient cardiology setting

Lida Apergi, Bruce Golden, John Baras, Kenneth Wood

This research tackles the problem of outpatient scheduling in the cardiology department of a large medical center, where patients have to go through a number of diagnostic tests and treatments before they are able to complete a final interventional procedure or surgery. The authors develop an integer programming (IP) formulation to ensure that the outpatients will go through the necessary procedures on time, that they will have enough time to recover after each step, and that their availability will be taken into account.

Operations Research for Health Care

2019

Integrating a safety smart list into the electronic health record decreases intensive care unit length of stay and cost

Daniel Lemkin, Benoit Stryckman, Joel Klein, Jason Custer, William Bame, Louis Maranda, Kenneth Wood, Courtney Paulson, Zachary Dezman

Integrating a safety smart list into the electronic health record decreases intensive care unit length of stay and cost.

Journal of Critical Care


Top