2015 Informs Annual Meeting

SD63

INFORMS Philadelphia – 2015

4 - Morning Commute Management Considering Commuters’ Aversion to Credit Loss Mohammad Miralinaghi, Purdue University, West Lafayette, IN, 47906, United States of America, mohammad.miralinaghi@gmail.com, Srinivas Peeta Under the tradable credit scheme, this study analyzes commuters’ departure time choices considering their aversion to credit loss. The analysis helps in determining credit price to manage morning commute congestion. The existence and uniqueness of the equilibrium credit price are investigated and a linear model is developed to obtain system optimum credit allocation and charging schemes. 5 - Effect of Glare on Shoulder-mounted Guide Sign Visibility Mohammed Obeidat, Kansas State University, Manhattan, KS, 66502, United States of America, moh2001ie@yahoo.com, Malgorzata Rys Glare is a serious concern in roadway safety during nighttime driving. Shoulder- mounted guide sign visibility will be evaluated under presence of glare in a field experiment using different retroreflective sheeting. Several variables will be considered. Data will be analyzed statistically to determine the significant variables that contribute to sign’s visibility. Nicholson Student Paper Competition II Cluster: Nicholson Student Paper Competition Invited Session Chair: Mark Squillante, IBM Research, Thomas J. Watson Research Center, 1101 Kitchawan Road, Yorktown Heights, NY, United States of America, mss@us.ibm.com 1 - Nicholson Student Paper Competition Illya Hicks, Rice University, 6100 Main MS-134, Houston, TX, 77005, United States of America, ivhicks@rice.edu This session highlights the finalists for the 2015 George Nicholson Student Paper Competition. Chair: Debarun Bhattacharjya, IBM T. J. Watson Research Center, 1101 Kitchawan Road, Rt. 134, Yorktown Heights, NY, 10598, United States of America, debarunb@us.ibm.com 1 - Balancing Research and Funding Needs: Value of Information and Portfolio Tools for Nano Risk Decisions Matthew Bates, Research Engineer, US Army Corps of Engineers, Engineer R&D Center, 696 Virginia Rd, Concord, MA, 01742, United States of America, Matthew.E.Bates@usace.army.mil, Jeffrey Keisler, Niels Zussblatt, Kenton Plourde, Ben Wender, Igor Linkov Nanotechnologies are economically and technically promising yet pose risks. Research may identify risks and paths to make technologies less hazardous or more acceptable. Given limited resources, funders need to prioritize research efforts. Current prioritization is done primarily thorough committee or executive decision. We apply value of information and portfolio analysis techniques to develop an efficient frontier of hazard research sets across three prominent nanomaterials (Ag, TiO2, MWCNTs). 2 - Scoring Rules, Value of Information, and Sensitivity Analysis Victor Richmond Jose, Georgetown University, 544 Hariri Bldg, 37th & O Sts NW, Washington, DC, United States of America, vrj2@georgetown.edu, Emanuele Borgonovo, Gordon Hazen, Elmar Plischke Scoring rules & value of information (VOI) are useful tools in decision analysis that measure the information content of data. In this talk we bridge these two seemingly separate areas of research. We obtain analytic expressions for VOI associated with some scoring rules and show that the resulting VOI sensitivity measures are global sensitivity measures that fall in a common rationale. We study this common rationale & obtain conditions that characterize properties of these sensitivity measures. SD63 63-Room 112B, CC SD64 64-Room 113A, CC Value of Information Sponsor: Decision Analysis Sponsored Session

3 - Valuing Data: A Closed Form Solution for the Expected Value of Sample Information Adam Fleischhacker, Assistant Professor Of Operations Management, University of Delaware, 222 Lerner Hall, Newark, DE, 19716, United States of America, ajf@udel.edu, Pak-wing Fok, Mokshay Madiman We present a method for valuing sample data prior to its collection. This valuation is given in closed-form and is flexible enough to mimic multiple decision making contexts. Compared to existing techniques, it provides tighter estimates of information value, insight as to the conditions under which data is valuable, and insight into the amount of data required to achieve certain levels of value. 4 - Preference Elicitation Schemes, Random Utility Models and the Value of Information Debarun Bhattacharjya, IBM T. J. Watson Research Center, 1101 Kitchawan Road, Rt. 134, Yorktown Heights, NY, 10598, United States of America, debarunb@us.ibm.com, Stephane Deparis Behavioral research indicates that when posed with preference elicitation queries, people provide inconsistent responses that depend on contextual factors. Random utility models have been proposed as a potential way to represent such inconsistencies. In this talk, we introduce a hierarchical Bayesian approach where the system is uncertain about the noise in a decision maker’s responses to queries, and present a methodology to compute the value of information from various elicitation schemes. SD65 65-Room 113B, CC Decision Analysis in Procurements and Procurement Auctions Sponsor: Decision Analysis Sponsored Session Chair: Janne Kettunen, Assistant Professor, The George Washington University, 2201 G Street, NW, Washington, DC, 20052, United States of America, jkettune@email.gwu.edu 1 - Evaluating Technology Readiness for Adoption and Integration at Navy Installations Eva Regnier, Associate Professor, Naval Postgraduate School, 699 Dyer Road, Monterey, CA, 93943, United States of America, eregnier@nps.edu, Robert Barron, Daniel Nussbaum While the DOD has adopted and adapted NASA’s technology readiness level (TRL) definitions in many contexts to measure its technologies’ progress, the achievement of high TRL levels has not necessarily led to adoption operationally. We propose a measure of the progress in removing specific barriers to technology adoption and integration in Navy installations that includes criteria for technical readiness, as well as stakeholder acceptance and removal of barriers to procurement. 2 - Scheduling Procurement Auctions Janne Kettunen, Assistant Professor, The George Washington University, 2201 G Street, NW, Washington, DC, 20052, United States of America, jkettune@email.gwu.edu, Young Kwak We derive conditions when scheduling procurement auctions impacts significantly on their expected costs. To help the procurement auction owner in scheduling the auctions, we develop a non-linear integer programming model, which we reformulate as a mixed integer programming problem to make it computationally amenable. We apply the model for the Florida Department of Transportation procurement auction data. Our results indicate that the optimal schedule can provide substantial cost savings. 3 - Decision Analysis Concepts in Public Procurement Jay Simon, American University, jaysimon@american.edu Decision analysis can play a significant role in helping carry out public procurement effectively. Public procurement decisions often require assessing preferences over multiple attributes. There may also be uncertainty regarding one or more elements of the process. Recent work has examined some of the unique challenges in public procurement, and how decision analysis techniques can be used to improve outcomes. 4 - On the Inefficiency of Multiattribute Auctions for Post-Auction Produced Goods Gregory Kersten, Prof, Concordia University, 1455 De Maisonneuve Blvd. W, 1450 Guy Street, Montreal, Qu, H3H 0A1, Canada, gregory.kersten@concordia.ca Multiattribute auctions are used to procure heterogeneous products. If they are produced post-auction, then the price and other attributes are interrelated and the assumption of the buyers’ and the sellers’ quasi-linearity utility does not hold. The implication is that while winning bids may be efficient solutions the auctions are inefficient mechanisms. The inefficiency and the possibility of improving the winning efficient bids via the side payment is illustrated with Cobb-Douglass economy.

140