Informs Annual Meeting 2017

MC78B

INFORMS Houston – 2017

4 - Friend or Foe? the Impact of Energy Storage on Solar and Wind Investments John Erik Bistline, Electric Power Research Institute, 3300 Sawtelle Blvd, Apt. 310, Los Angeles, CA, 90066, United States, john.bistline@gmail.com Energy storage is often linked with improving the economic proposition of variable renewables like wind and solar, though few existing studies examine how low-cost storage simultaneously impacts electric sector operations and investments. This talk investigates the economics of grid-connected energy storage and its effect on system investments, emissions, and costs using the US- REGEN capacity planning and dispatch optimization model. Model experiments illustrate how the effects of storage depend on the supply responsiveness of different generators and vary across different market conditions, which may favorably or adversely influence solar and wind deployment. 5 - A Future Outlook on Wind, Solar, and Energy Storage Deployment Cara Marcy, Renewable Electricity Analyst, U.S. Energy Information Administration, Washington, DC, 20585, United States, cara.marcy@eia.gov EIA’s National Energy Modelling System (NEMS) provides a future multi-decadal assessment of U.S. energy sectors, including electricity. Recently, variable technologies, such as wind and solar, have increased in market share. In NEMS, challenges arise in accounting for this variability in a manner that preserves computational time. In addition energy storage is one option that can take advantage of value streams presented from this variability, such as curtailments, as well as other market value streams. This presentation will review the model structure for energy storage and curtailments in NEMS, as well as highlight scenarios that study the relationship between renewables and energy storage. 380B Humanitarian Logistics Contributed Session Chair: Shreya Gupta, The University of Texas-Austin, Austin, TX, United States, shreya.gupta@utexas.edu 1 - Exact and Heuristic Approaches for Bidirectional Leaf Trajectory Optimization in Intensity-modulated Radiotherapy Seyedali Mirzapour, Wichita State University, 203A-Engineering Building, 1845 Fairmount St,, Wichita, KS, 67260, United States, mirzapour.ie@gmail.com, Ehsan Salari Traditionally, unidirectional leaf-sweeping schemes have been used to dynamically modulate radiation beams in radiotherapy. However, the unidirectional leaf-motion restriction may lead to poor beam modulation when the available beam-on time is limited. This research investigates the potential gain in beam-modulation quality and delivery efficiency obtained from relaxing this restriction and allowing for bidirectional leaf trajectories. A mixed-integer programming formulation with exact and heuristic solution methods are developed to optimize for bidirectional piecewise-linear leaf trajectories. 2 - Optimal Child Delivery Strategies for Hypertensive Disorders of Pregnancy Aysegul Demirtas, PhD Candidate, Arizona State University, Tempe, AZ, United States, Aysegul.Demirtas@asu.edu, Esma S. Gel, Soroush Saghafian, Dean V. Coonrod Hypertensive disorders of pregnancy (HDP) constitute one of the leading causes of maternal and neonatal mortality and morbidity. We study the decision problem of timing of child delivery for women with HDP. We formulate a Markov decision process (MDP) model that minimizes the risks of maternal and neonatal adverse outcomes, and assess its results with a probabilistic sensitivity analysis (PSA). In PSA, we construct instances by generating model parameters in restricted orders to incorporate physicians’ intuition in part of the estimations with low sample size. We also build a robust MDP model that considers the sensitivity of transition probabilities while avoiding over-conservative policies. 3 - Analytic Advancements in the Project Data Sphere Open Access Data Sharing Platform Steven B.Cohen, Vice President, RT.I.International, 701 13th Street NW, Washington, DC, 20005-3967, United States, scohen@rti.org Project Data Sphere (PDS) is a research platform providing broad access to both de-identified patient-level data from oncology clinical trials and related analytic tools. Data providers are required to de-identify patient-level data by removing key demographic data. To address these analytic constraints, the patient-level cancer phase III clinical data has been augmented by linking the social, economic and health related characteristics of like cancer survivors from nationally representative health and healthcare-related survey data. This presentation will provide an overview of the methodologies used and an analysis of the association between these factors and patient outcomes. MC78B

4 - A Recommendation System for Managing Physical Fatigue in the Workplace Lin Lu, Auburn University, 371 W. Glenn Ave, Apt 13, Auburn, AL, 36830, United States, lzl0032@auburn.edu, Fadel Mounir Megahed, Lora Cavuoto In this talk, we propose a system for managing fatigue in the workplace based on combining principles from artificial intelligence, safety science and management. Our system consists of the following modules: (a) a knowledge base, (b) an inference engine, and (c) a user interface. The system recommends appropriate interventions based on the type of fatigue and the different constraints at the workplace. 5 - Using Decision Analytics to Provide Patient-centered Care in Battling Childhood Epilepsy Shreya Gupta, PhD Student, The University of Texas-Austin, 3452 Lake Austin Blvd, Apt E, Austin, TX, 78703, United States, shreya.gupta@utexas.edu, John Hasenbein We propose using decision analytics methods from operations research and statistics to improve quality of life and provide patient centered care in childhood epilepsy. Epilepsy researchers estimate that at least 90,000 children per year are not receiving the epilepsy treatment they need, leading to years of poor quality of life and increased mortality for these children. We propose building a Markov decision process (MDP) model of the decision process for childhood epilepsy, utilizing data analytics and personal utility analysis to provide decision support systems for both providers and patients. The primary advantage of such models is that they can be tuned for each patient and/or provider. 381B Uncertainty-aware Expansion Planning in Complex Infrastructures Sponsored: Energy, Natural Res & the Environment Electricity Sponsored Session Chair: Yuri Dvorkin, New York University, New York, NY, 11201, United States, dvorkin@nyu.edu 1 - The Value of Demand Response Controllability Kenneth Bruninx, KU. Leuven, Celestijnenlaan 300, Post Box 2421, Leuven, B3001, Belgium, Kenneth Bruninx, EnergyVille, Genk, Belgium, Kenneth Bruninx, VITO, Mol, Belgium, kenneth.bruninx@mech.kuleuven.be Demand response (DR) resources, such as thermostatically controlled loads, are not fully controlled by the system operator and their availability is limited by user-defined comfort constraints. We present an operational model that optimizes power plant scheduling in the presence of stochastic RES-based generation and DR resources (here: residential electric heating systems) that are partially controllable and governed by user-defined comfort constraints. This model is used to evaluate the operating cost savings that can be attained via DR on a model of the Belgian power system. 2 - Data-driven Stochastic Transmission Expansion Planning Jianhui Wang, Southern Methodist University, Dallas, TX, United States, jianhui@smu.edu Jianhui Wang, Argonne National Laboratory, Argonne, IL, United States, jianhui@smu.edu We present a data-driven two-stage stochastic transmission expansion planning method with uncertainties. We first use a confidence set for the unknown distribution of the uncertain parameters. Then, a two-stage data-driven transmission expansion framework is developed considering the worst-case distribution within the constructed confidence set. A decomposition framework embedded with Benders’ and Column-and-Constraint generation methods is used to solve the problem. 3 - Optimization of Integrated Gas-electric Systems under Uncertainty Line Roald, Los Alamos National Laboratory, Los Alamos, NM, United States, lineroald@gmail.com, Sidhant Misra The electric grid depends on gas-fired generation to balance the intermittent renewable energy production, resulting in time-varying and unpredictable gas flows. To ensure secure and economic operation of the integrated system, we solve an optimization problem that minimizes cost of operation, subject to constraints from both the electric and gas systems. The electric grid is modelled using chance constraints, which provide probabilistic bounds on the gas withdrawals. The gas constraints are modelled using partial differential equations, and are formulated as robust constraints based on monotonicity properties of the gas flows. We demonstrate the benefits of the method through a case study. MC79

232