2015 Informs Annual Meeting

MA67

INFORMS Philadelphia – 2015

MA66 66-Room 113C, CC Air Traffic Management and Airline Operations Sponsor: Aviation Applications Sponsored Session Chair: Peng Wei, Assistant Professor, Iowa State University, Aerospace Engineering Department, 2312 Howe Hall, Ames, IA, 50011, United States of America, pwei@iastate.edu 1 - Modeling Airport Operations for Capacity Planning by Airport Planners and Individual Carriers L. Douglas Smith, University of Missouri-St. Louis, St. Louis, MO, United States of America, ldsmith@umsl.edu, Jan Fabian Ehmke, Deng Pan, Liang (leon) Xu, Ziyi Wang We discuss a blend of statistical modelling, LP optimization and discrete-event simulation that was used in the construction, calibration, validation and application of a model for capacity planning at commercial airports. With results from simulated scenarios, we examine the effects of changing usage of runways, taxiways, ramps and gates under different traffic conditions. We also see how concentration of activity at hub airports can cause flight delays to cascade for individual carriers. 2 - Real-time Gate Assignment Diego Klabjan, Northwestern University, Evanston, IL 60208, United States of America, Professor, d-klabjan@northwestern.edu Due to unpredictable events airlines have to frequently adjust gate assignments. We propose a network flow model that also considers crew and passenger connections. A thorough computational study is presented based on data from a big US airport. 3 - An Alternative Approach to Capacity Allocation at Congested Airports Amedeo Odoni, Professor, MIT, Room 33-219, MIT, 77 Mass Ave, Cambridge, MA, 02139, United States of America, arodoni@mit.edu, Alexandre Jacquillat Current slot allocation at congested airports worldwide is based on long-standing guidelines developed by IATA. An alternative approach, developed after extensive research, will be outlined: it is more responsive to airline preferences, integrates airport operating capabilities, minimizes interference with airline competitive scheduling and is based on achieving user-specified performance goals, instead of being solely driven by arbitrary and administratively-determined capacity constraints. 4 - Airline Passenger Origin-Destination Reaccommodation with Spare Aircraft Peng Wei, Assistant Professor, Iowa State University, Aerospace Engineering Department, 2312 Howe Hall, Ames, IA, 50011, United States of America, pwei@iastate.edu The researchers at Amadeus have presented a method to formulate airline passenger origin-destination reaccommodation into a multi-commodity flow problem, for which they also designed the solution algorithm. In this paper we consider the Airline Passenger Origin-destination Reaccommodation with Spare Aircraft, which no one has tackled before. MA67 67-Room 201A, CC City Logistics and Sustainable Urban Freight Systems - II Sponsor: TSL/Freight Transportation & Logistics Sponsored Session Chair: Johanna Amaya, Rensselaer Polytechnic Institute, 110 8th St. JEC 4037, Troy, NY, 12180, United States of America, amayaj@rpi.edu 1 - Public Sector Initiatives Sustainable Urban Freight Systems Part I Johanna Amaya, Rensselaer Polytechnic Institute, 110 8th St. JEC 4037, Troy, NY, 12180, United States of America, amayaj@rpi.edu, Jose Holguin-Veras, Carlos A. Gonzalez-Calderon, Stacey Hodge, Michael Browne, Miguel Jaller, Jeffrey Wojtowicz, Cara (xiaokun) Wang, Daniel Haake, Ivan Sanchez-Diaz Transportation policy should ensure freight is moved efficiently as it is a physical manifestation of the economy. In this study, planners are provided with guidelines to implement initiatives addressing freight issues. Initiatives were organized as part of a continuum, from supply to demand side; underpinned by stakeholder engagement. This paper discusses pricing, incentives, taxation; logistical management; and freight demand/land use, along with a methodology to identify potential initiatives

3 - Global Uncertainties in Pharma R&D Portfolio Management Chris Dalton, Syncopation Software, 6 State Street, Suite 308, Bangor, ME, 04401, United States of America, cdalton@syncopation.com R&D portfolio management in pharma tends to be focused on individual project risk factors, technical and commercial. Global risk factors affecting many projects are usually ignored, on the theory that while they may change the value of the portfolio they are unlikely to alter the project ranking. The possibility of far reaching changes in government policy suggests that this assumption should be tested. This talk examines the impact of some potential policy changes on a notional portfolio. 4 - Manufacturing Investments in Pharmaceutical R&D: Should I Pull the Trigger? Phil Beccue, White Deer Partners, Westlake Village, CA, United States of America,phil@beccue.com Capital investments to manufacture biopharmaceuticals require long lead times and large expenditures that require board approval. These facilities produce multiple products in highly competitive markets with uncertain demand, and if not planned carefully, can result in a loss of economic value or inefficient allocation of scarce resources. Based on recent case studies, we will explore how decision analytics helps guide manufacturing choices. MA65 65-Room 113B, CC From Value of Information to Sensitivity Analysis Sponsor: Decision Analysis Sponsored Session Chair: Emanuele Borgonovo, Professor, Bocconi University, Via Roentgen 1, Milan, 20136, Italy, emanuele.borgonovo@unibocconi.it 1 - Importance Measures and Value of Information Alessandra Cillo, Assistant Professor, Bocconi University, Via Roentgen 1, Milan, Italy, alessandra.cillo@unibocconi.it, Emanuele Borgonovo The importance measures are sensitivity indices. Despite their wide use, no work has systematically addressed the relationships among them. We provide new probabilistic relationships that link each measure to the remaining ones. These new relationships also allow us to introduce new importance measures, value of information based. We compare an importance measure and the value of information-based measure, which depends upon the same measure: interesting differences emerge. 2 - Integral Sensitivity in Linear Programming Richard Wendell, Professor, University of Pittsburgh, Katz Graduate School of Business, Pittsburgh, PA, 15238, United States of America, wendell@katz.pitt.edu, Emanuele Borgonovo, Greg Buzzard Herein we provide a general framework merging two well known sensitivity analysis approaches, namely, the tolerance sensitivity and global sensitivity approaches. This unification leads to a series of novel results, ranging from analytical formulae for computing global sensitivity measures to conditions ensuring that uncertainty in the parameters is immaterial. Moreover, the results provide the basis for a numerical approach for assessing key problem drivers and determining tolerance regions. 3 - Robust Importance Meaures Emanuele Borgonovo, Professor, Bocconi University,

Via Roentgen 1, Milan, 20136, Italy, emanuele.borgonovo@unibocconi.it

Importance measures are probabilistic indicators that deliver information about key risk drivers. They are defined for aleatory uncertainty. We show that, relying on their probabilistic meaning, one can obtain a natural extension of importance measures also when probabilities are not known.

167