Polynomial Function Reformulation via Mixed-Integer Optimization
This study presents a new mathematical reformulation for solving polynomial integer nonlinear optimization (PINLO) problems, which linearizes polynomial functions of bounded integer variables of any degree. Through computational experiments, the study demonstrates that these integer linear optimization (ILO) reformulations can efficiently handle large-scale PINLO problems using Gurobi, with capabilities exceeding those of current leading solvers like BARON, highlighting its potential for real-world applications.
Area: Methodology
Collaborators
Pitchaya Wiratchotisatian
Andrew Trapp
Refugee Camp Management
This study addresses the challenge of distributing aid effectively to both camp-based and urban refugees, amidst administration difficulties, demand uncertainty, and funding volatility. The policy evaluates costs related to redirecting urban refugees, insufficient aid for camp-based refugees, and excess inventory. The study offers valuable insights for managing humanitarian aid allocation under uncertain conditions.
Area: Immigration
Collaborators
Shima Azizi
Cem Deniz Caglar Bozkir
Andrew Trapp
O. Erhun Kundakcioglu
Ali Kaan Kurbanzade
Sponsors
NSF CMMI-1825348
Human-Centric Decision Support Tools
This book chapter discusses the role of well-designed decision support tools in enhancing complex decision-making processes. The chapter advocates for a deep understanding of stakeholder needs to build trust and ensure successful tool adoption, while also cautioning against the pitfalls of poorly designed tools. Insights and best practices from various examples are shared to guide the development of effective decision support tools.
Area: Methodology
Collaborators
Narges Ahani
Andrew Trapp
Sponsors
NSF CMMI-1825348
Dynamic Course Scheduling
This study introduces a novel course scheduling framework for universities to navigate strategic university scaling in the long term. Our optimization-driven framework includes utilization-related objectives, aiming to optimize space management. Through extensive experimentation, we show tradeoffs between additional students and associated costs and discuss different strategies to address bottleneck course sections. The study provides valuable insights for universities to make long-term strategic decisions on investments while accounting for student body expansion or contraction.
Area: Methodology
Collaborators
Özge Aygül
Teodor Hellgren
Shima Azizi
Andrew Trapp
Sponsors
WPI
Improving Monitoring of Border Transit Stations Using Performance Analytics
This project involves the development of a data envelopment analysis (DEA) system in collaboration with Love Justice International (LJI), an NGO active in anti-human trafficking efforts. The system is designed to evaluate the performance of transit monitoring stations along the Nepal-India border, a method aimed at identifying potential human trafficking cases preemptively. To the best of our knowledge, this is the first application of DEA in the anti-human trafficking domain
Area: Anti-Human Trafficking
Collaborators
Geri Louise Dimas
Malak El Khalkhali
Alex Bender
Jeffrey S Blom
Renata Konrad
Kayse Lee Maass
Andrew Trapp
Joe Zhu
Sponsors
NSF CMMI-1841893
Operations and Analytics for Improving Immigration
This study features a systematic literature review on the current state of research that applies Advanced Analytics, such as Operations Research and Data Science, within the domain of immigration operations. We particularly explore how such advancements can contribute to societal well-being by promoting more efficient, equitable, and humane immigration processes.
Area: Immigration
Collaborators
Marcela Vasconcellos
Fatemeh Farajzadeh
Geri Louise Dimas
Andrew Trapp
Sponsors
NSF CMMI-1935602
Improving US Defensive Asylum Process
The United States immigration court system faces a significant backlog, with nearly 1.5 million cases awaiting hearings, causing delays that strain both government and community resources. This project employs discrete event simulation to analyze and deconstruct the complex elements of the immigration court system, aiming to improve efficiency and reduce delays. By simulating and adjusting factors like case assignments, queuing disciplines, and priority queues, the project offers insights into data-driven solutions for streamlining the court process and addressing the backlog effectively.
Area: Immigration
Collaborators
Geri Louise Dimas
Renata Konrad
Andrew Trapp
Sponsors
NSF CMMI-1825348
Border Operations Management
This book chapter reviews the operational challenges arise with large movements of human flows toward national borders. These challenges constitute a complex humanitarian crisis that requires appropriate preparations of border resources. Motivated by prevailing national agenda issues for improved border support, common operational challenges along the border are identified. This book chapter highlights how optimization can be employed to improve both the security and humanitarian aspects of border operations in the context of migration flows.
Area: Immigration
Collaborators
Fatemeh Farajzadeh
Andrew Trapp
Sponsors
NSF CMMI-1825348
Optimizing the Foster Care Visitation Scheduling Problem
This project addresses the scheduling challenges faced by child welfare agencies in ensuring regular foster child visitations for a county partner in New York State. Visitation scheduling can be challenging due to fluctuating caseloads and fixed workforce levels. Mathematical optimization is employed, including advanced network optimization approaches, to identify optimal schedules and routes for county workers to transport foster children to visitation meetings. The methods are integrated into a web-based interface for improved operational efficiency and visit consistency in foster care.
Area: Foster Care
Collaborators
Shima Azizi
Caroline Johnston
Enes Tanrıkulu
O. Erhun Kundakcioglu
Andrew Trapp
Locating Residential Human Trafficking Shelters
This study addresses the significant role of residential shelters in aiding the stabilization and reintegration of trafficked persons into society. Using concepts from health and social welfare economics, we develop an optimization model that allocates a budget for locating residential shelters in a manner that maximizes a measure of societal impact while respecting budgetary constraints. We illustrate the utility of the model via our case study that allocates a budget among a candidate set of residential shelters for female sex trafficking survivors in the United States.
Area: Anti-Human Trafficking
Collaborators
Kayse Lee Maass
Andrew Trapp
Renata Konrad
Optimizing Sponsored Humanitarian Parole
This study focuses on a novel approach to assist Ukrainian refugees seeking humanitarian parole in the United States following Russia's invasion of Ukraine in 2022. RUTH (Refugees Uniting Through HIAS) is a software that implements the Thakral Multiple-Waitlist Procedure for the first time in refugee resettlement history to match refugees to host communities based on refugees' locational preferences and sponsors' priorities. This research suggests that such systems could greatly enhance the efficiency and equitability of other rapidly deployed humanitarian parole processes, offering a template for future initiatives.
Area: Immigration
Collaborators
Fatemeh Farajzadeh
Ryan Killea
Alexander Teytelboym
Andrew Trapp
Sponsors
HIAS; NSF CMMI-2233377
Improving Neighborhood Support Teams Operations
This study provides a multi-criteria performance analysis of Neighborhood Support Teams (NSTs) and their role in aiding Afghan refugee families who resettled in the US as humanitarian parolees following the Afghan crisis of Fall 2021. The NSTs, coordinated by Ascentria Care Alliance, consist of diverse groups dedicated to supporting these refugees. The study uses a community-based participatory action research (CBPAR) methodology, encompassing two phases aimed at evaluating the effectiveness of NSTs and identifying best practices for assisting new arrivals in the US.
Area: Immigration
Collaborators
Fatemeh Farajzadeh
Teodor Hellgren
Sarah Stanlick
Andrew Trapp
Sponsors
Ascentria Care Alliance; NSF CMMI-2233377
Understanding Factors that Drive Better Child Welfare Operations
Public child welfare agencies play a pivotal role in safeguarding the well-being of children and thus, the future of our society. This study conducts a multi-criteria analysis for benchmarking the performance of the United States child welfare system. Our study offers data-driven directions for child welfare agencies to improve safety and permanency outcomes for children.
Area: Foster Care
Collaborators
Sepideh Sedghi
Shima Azizi
Andrew Trapp
Border Operations Management
This study addresses the escalating challenge of international migration, highlighting that in 2023, the number of people forcibly displaced has surpassed 100 million, a 13% increase from the previous year. Stochastic programming is used to guide cost-effective decisions for locating processing facilities and pre-allocating critical and scarce support resources at national borders. The framework can enhance strategic and operational resource allocation for managing migrant flows at international borders, offering a proactive approach to a pressing global issue.
Area: Immigration
Collaborators
Fatemeh Farajzadeh
Rashika Jakhmola
Luke Caddell
Andrew Trapp
Sponsors
NSF CMMI-1825348
Optimizing Community Paramedicine Service Delivery
This study introduces an optimization framework for enhancing the efficiency of community paramedicine programs in the United States. The model introduced in this study aims to enhance patient welfare, reduce hospital costs, and lower readmission and emergency department visits.Using real data from a hospital system in Upstate New York, the study conducts computational experiments to test the framework. The results demonstrate the model's capability to offer promising insights for managing and improving community paramedicine programs.
Area: Healthcare
Collaborators
Shima Azizi
Özge Aygül
Brenton Faber
Sharon Johnson
Renata Konrad
Andrew Trapp
Sponsors
Healthcare Delivery Institute
Optimizing Refugee Resettlement with Advanced Analytics
This project is designed to improve the initial placement of refugees in host countries for the first time by using advanced analytics. Annie™ MOORE (Matching and Outcome Optimization for Refugee Empowerment) stands as the first of its kind, a software empowered by integer programming and data analytics to aid resettlement agencies in making such decisions. Currently implemented and in use at HIAS, one of the nine non-profit organizations in the U.S. working with the Department of State. Further enhancements in this research include the dynamic approach to placement and the uncertainties in match quality scores by incorporating a family-level risk aversion strategy.
Area: Immigration
Collaborators
Narges Ahani
Paul Gölz
Ariel D. Procaccia
Tommy Andersson
Alessandro Martinello
Alexander Teytelboym
Fatemeh Farajzadeh
Osman Özaltın
Andrew Trapp
Sponsors
NSF CMMI-1825348
Runaway and Homeless Youth Shelter Capacity Expansion
This study addresses the critical issue of runaway and homeless youth and young adults (RHY) in the United States. The research focuses on New York City and adopts a data-driven methodology to estimate the collective capacity required by service providers to meet the needs of RHY adequately, including those most at risk of being trafficked. The proposed integer programming model is informed by partnerships with key stakeholders and is designed to accommodate various complexities such as time-dependent allocation, capacity expansion, stochastic youth arrivals, variable lengths of stay, periodically provided services, and specific service delivery time windows.
Area: Anti-Human Trafficking
Collaborators
Yaren B. Kaya
Kayse Lee Maass
Geri Louise Dimas
Renata Konrad
Andrew Trapp
Meredith Dank
Optimizing Benefit to Cost Ratio in Public Sector
This study presents a new framework designed to enhance decision-making in the public sector, particularly focusing on maximizing the benefit to cost ratio (BCR) for public sector decisions.
Area: Homelessness
Collaborators
Forrest Miller
Yaren B. Kaya
Geri Louise Dimas
Renata Konrad
Kayse Lee Maass
Andrew Trapp
Sponsors
NSF CMMI-1935602
OR and Analytics for Combatting Human Trafficking
Human trafficking, a serious global issue affecting social, economic, and human rights dimensions, is being increasingly studied within the Operations Research (OR) and Analytics fields. This project systematically reviews the growing body of research focusing on diverse methodologies and theoretical approaches, underscoring the collective impact of these fields in anti-trafficking efforts. This body of work collectively illustrates the critical role of OR and Analytics in tackling the complex, multifaceted issues surrounding human trafficking, guiding future research towards more effective prevention and intervention strategies.
Area: Anti-Human Trafficking
Collaborators
Renata Konrad
Geri Louise Dimas
Kayse Lee Maass
Andrew Trapp
Timothy Palmbach
Jeffrey S Blom
Sponsors
NSF CMMI-1841893, CMMI-1935602
Many-To-One Stable Matching
This study presents integer optimization models for stable many-to-one matching problems, addressing issues like incomplete preference lists and ties. It introduces new constraint sets for preventing envy and waste, along with algorithms for faster constraint generation. It also proposes aggregate objective functions with hierarchical prioritiesThis approach highlights the adaptability and efficiency of optimization-based methods in complex matching scenarios.
Area: Methodology
Collaborators
Pitchaya Wiratchotisatian
Hoda Atef Yekta
Andrew Trapp
Many-To-Many Stable Matching for Migrant Employment
This study focuses on securing stable employment for vulnerable persons through an advanced platform using mathematical optimization and AI. Concepts such as many-to-many matching, preference-based matching, and bias reduction are explored. The results of this project, deployed in the real world, promise improved and fairer outcomes for refugees.
Area: Immigration
Collaborators
Marcela Vasconcellos
Andrew Trapp
Roee Shraga
Swati Gupta
Sponsors
Anonymous Donor
Nonprofit Resource Sharings
This project addresses the resource challenges faced by nonprofit organizations (NPOs) by introducing SWAP, a novel resource-sharing system. SWAP allows NPOs to exchange resources through a collaborative auction-based process, using the virtual currency SWAPcredit for liquidity. The system includes a central technology that optimizes resource exchanges, and an online platform, SWAP Hub, for managing offers and bids. A human-centric co-design approach ensures practical solutions shaped by NPO professionals. Implemented in Howard County, Maryland, SWAP demonstrates strong potential for broader expansion.
Area: Nonprofit Operations
Collaborators
Weixiao Huang
Elise Deshusses
Jennifer A. Pazour
Arjun Venat
Yunus D. Telliel
Sarah Stanlick
Andrew Trapp
Sponsors
NSF FW-HTF-2222713, FW-HTF-2222697
Cohort Formation in SWAP
With the successful deployment of the nonprofit resource sharing platform SWAP, there is an associated challenge to create cohorts for resource sharing among nonprofits to ensure positive exchange experiences. The project identifies key attributes for resource sharing and uses them in an integer optimization to generate optimal cohort formations. Experiments using both real and simulated data validate the effectiveness of the model, providing a roadmap for forming successful nonprofit cohorts.
Area: Nonprofit Operations
Collaborators
Weixiao Huang
Andrew Trapp
Sarah Stanlick
Sponsors
NSF FW-HTF-2222713, FW-HTF-2222697
Alternative Approaches to Solve Mixed-Integer Nonlinear Optimization Problems
This project presents an innovative approach to solving nonlinear discrete optimization problems with unknown resource vectors. Using a tree-based data structure, we efficiently handle queries and rapidly analyze regions of interest by constructing level-sets of the value function. The method integrates problem structure with incremental solution construction, yielding a scalable algorithm with promising results in computational experiments.
Area: Methodology
Collaborators
Ryan Killea
Junlong Zhang
Osman Özaltın
Andrew Trapp