Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MB43

2 - Lives Saved vs. Lives Lost in Survey Research: Investigating Methodological Consistency Candice Huynh, California State Polytechnic University, Pomona, College of Business, TOM, 3801 W. Temple Ave, Pomona, CA, 91768, United States, Jeffery L. Guyse, L. Robin Keller Seven different elicitation procedures are employed in a between-subjects experiment over hypothetical scenarios involving lives either being saved or lost over time. Inspired by a study by Frederick (2003) which included lives being saved in 6 of the 7 procedures (choice, matching, total, sequence, equity, & context) and lives lost in just one of the cases (rating), we incorporate a completely balanced and symmetrical design with both lives saved and lives lost for all 7 methodologies. The effect on the survey responses due to the outcome being saved or lost along with the contextual effect of the questionnaire design itself are both analyzed for each of the 7 elicitation procedures employed. 3 - Psychophysics of Terror Attack Consequences Matthew Baucum, University of Southern California, 12540 Braddock Drive, Apt 105, Los Angeles, CA, 90066, United States, Heather Rosoff, Richard S. John We conceptualize public perception of the consequences of severe biological and radiological terror attacks along three dimensions: size of spatial area impacted, length of time of impact, and fatalities. We conducted an experiment in which respondents (n=384) read a vivid scenario about either a biological or radiological terror attack, varying in terms of spatial area impacted, length of time of impact, and fatalities. We develop psychophysical curves for all 3 consequence dimensions in terms of perceived attack severity, perceived risk, and reported fear. These curves are markedly non-linear, and in some cases, non-monotonic. 4 - The Effects of Assigning Buddies on Social Support in Online Communities Ali Esmaeeli, University of California, Irvine, 1914 Verano Place, Irvine, CA, 92617, United States, Cornelia Pechmann We look at the effects of assigning buddies on the social support that participants receive in an online community. We show that an active buddy provides more social support than another active community member. We also show that an inactive buddy doesn’t decrease a person’s engagement with others in the community. We develop four continuous and quantitative tie-strength measures to evaluate the level of participants’ engagement with others in online communities. These measures are quite general and can be used to evaluate tie- strength in other online communities. Sponsored Session Chair: Wie Xie, RPI 1 - An Envelope Procedure for Ranking and Selection Problems Shane Henderson, Cornell University, Ithaca, NY, 14853, United States, Sijia Ma We present the envelope procedure for solving ranking and selection problems, aka exploratory bandit problems. We assume that each system (or arm) can be repeatedly sampled independently, and returns a subGaussian random variable. We prove that the envelope procedure returns a selected system that is within a given distance (in mean) of the best system with given probability. We also present a complexity analysis and computational results showing that the envelope procedure has impressive empirical performance under a specific sampling rule. 2 - Distance Correlation Screening for Hierarchically Structured Data Ilya Ryzhov, University of Maryland, College Park, MD, 20742, United States, Yimei Fan, Yuan Liao In many applications of business and marketing analytics, common characteristics of products and customers are represented as categorical variables, and each category can be split up into multiple subcategories at a lower level of the hierarchy. Hundreds of thousands of binary variables may be used to model the hierarchy, which makes it very useful to screen out large numbers of irrelevant or insignificant variables. We propose a new dynamic screening method that screens out large parts of the hierarchy at higher levels, greatly reducing the computational burden of screening and producing much more powerful predictive models. This approach may be useful in other areas, such as simulation analytics. n MB42 North Bldg 227A Simulation Optimization I Sponsored: Simulation

3 - A Metamodel-assisted Framework for Two-stage Optimization via Simulation Wei Xie, Rensselaer Polytechnic Institute, 110 8th Street, CII, Room 5207, Troy, NY, 12180, United States, Yuan Yi For the discrete two-stage optimization with the unknown response obtained from simulation, we introduce a metamodel-assisted framework that can efficiently employ the simulation resource to iteratively solve for the optimal first- and second-stage decisions. At each visited first-stage decision, we develop a local metamodel to solve a set of deterministic recourse problems simultaneously. Then, we construct a global metamodel accounting for the finite sampling error from SAA and the second-stage optimality gap. Assisted by this global-local metamodel, we propose a simulation optimization approach that can efficiently guide the search for the optimal first- and second-stage decisions. n MB43 North Bldg 227B Emerging Topics in Energy Systems Integration Emerging Topic: Energy and Climate Emerging Topic Session Chair: Dharik Mallapragada, ExxonMobil Corporate Strategic Research, ExxonMobil Corporate Strategic Research, Annandale, NJ, United States 1 - Evaluating Tradeoffs in Modeling Approaches for Electric Power Systems Analysis Nidhi Santen, Electric Power Research Institute, Palo Alto, CA, United States, John Bistline Throughout electric power systems analysis, simplifying assumptions and approaches are required to make models computationally tractable while still representing salient features of the system. The objective of this research is to test and examine the benefits of alternate modeling approaches in four key areas: temporal resolution, spatial resolution, representation of end use, and representation of uncertainty. We calculate and examine potential impacts of using simplified methodologies in these areas via a set of exercises using the U.S. long-range electric generation capacity planning model, US-REGEN. 2 - Emerging Topics in Energy Systems Integration Vehicle Electrification Matteo Muratori, National Renewable Energy Laboratory From an energy perspective, transportation is the least diversified demand sector, and offers numerous untapped opportunities for electrification over the next decades. While petroleum is still the major fuel used in transportation, adoption of electric vehicles in the light-duty vehicle market is increasing and electrification of other sub-sectors is gaining more interest. Vehicle electrification is creating new connections between the transportation and the electric sectors and coordinated charging is the greatest near-term opportunity for synergistic integration of electric vehicles in the power system and exploit vehicle flexibility to support grid operation. 3 - Impact of Model Resolution on Scenario Outcomes for Electricity Sector System Expansion Dharik Mallapragada, ExxonMobil Research and Engineering, 1545 Route 22 East, Annandale, NJ, 08801, United States, Dimitri Papageorgiou, Aranya Venkatesh Power sector capacity expansion models (CEMs) assess how techno-economic and policy drivers impact the cost-optimal generation capacity mix and their utilization over decadal time scales. Here, we evaluate the impact of incorporating operational detail in a CEM on the resulting outputs for various Texas grid scenarios. In general, we find that a traditional CEM with aggregated time blocks, overstates solar capacity and understates wind and natural gas capacity compared to an alternate CEM with chronological time-representation of grid operations. The findings imply the need for embedding sufficient temporal resolution and chronology in CEMs and broadly multi-sector energy-economic models 4 - The Role of Firm Low-carbon Electricity Resources in Deep Decarbonization of Power Generation Nestor A. Sepulveda, Massachusetts Institute of Technology, Cambridge, MA, 02142, United States We use a detailed power system investment and operations model to investigate the role of firm low-carbon resources in decarbonizing power generation in combination with variable renewable resources, energy storage, demand flexibility, and long- distance transmission interconnections. To increase the robustness of our results, we evaluate nearly 1000 cases covering a range of emissions limits, regional differences in demand and renewable resource potential, and a wide range of technology and cost uncertainties.

163