2015 Informs Annual Meeting

SB54

INFORMS Philadelphia – 2015

SB54 54-Room 108A, CC Pricing Inspired by Data and Practice

1 - A New Formulation for Energy-efficient Aggregation of Virtual Machines in Cloud Data Centers Hajime Miyazawa, Nanzan University, 18 Yamazato-cho, Showa-ku, Nagoya, 4668673, Japan, miyazawa@nanzan-u.ac.jp, Mihiro Sasaki One of the main concerns of cloud computing is energy consumption in data centers. Appropriate aggregation of virtual machines(VMs), which are the computing entities of cloud computing, on limited number of physical servers in data centers can reduce energy consumption by shutting down the rest of the servers. We present a new formulation of assigning VMs among physical servers to achieve energy-efficient VM aggregation in cloud data centers. 2 - Mathematical Properties of New Indices for Evaluating Spatial Demand-and-supply Balance Takamori Ukai, Tokai University, 143 Shimokasuya, Isehara, 2591193, Japan, ukai@tsc.u-tokai.ac.jp, Mihiro Sasaki In this presentation, we show some mathematical properties of new indices for evaluating spatial demand-and-supply balance. More precisely, we show that the sequence of solutions generated by the iterative algorithm converges to the proposed indices. We also discuss the relationship between the solutions and those obtained by solving a mathematical programming problem with an objective of minimizing the variance of indices. The objective corresponds to minimize unfairness among customers. 3 - Optimal Location Model for Anti-piracy Activity in Somalia Daisuke Watanabe, Associate Professor, Tokyo University of Marine Science and Technology, 2-1-6 Etchujima, Koto-ku, Tokyo, 135-8533, Japan, daisuke@kaiyodai.ac.jp, Richard Church A significant number of pirate attacks have occurred off the coast of Somalia. This is a major threat to ships navigating the major sea lane between Asia and Europe. The purpose of this study is to analyze the optimal location for anti-piracy activity in Somalia using the maximal covering location model. 4 - Optimal Real-time Pricing of Electricity for Supply and Demand Control in a Smart Community

Cluster: Tutorials Invited Session Chair: Georgia Perakis, MIT, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States of America, georgiap@mit.edu 1 - Tutorial: How Analytics Can Impact Promotion Pricing Georgia Perakis, MIT, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States of America, georgiap@mit.edu, Lennart Baardman, Maxime Cohen, Swati Gupta, Jeremy Kalas, Zachary Leung, Danny Segev, Kiran Panchamgam, Anthony Smith Pricing has seen exciting developments in the recent years. A particular area of pricing that has recently emerged is promotion pricing. In many important settings such as in grocery retail, promotions are a key instrument for driving sales. The Promotion Optimization Problem is a challenging problem as the retailer needs to decide which products to promote, what is the depth of price discounts, when to schedule the promotions and how to promote the product. We will discuss our collaboration with Oracle RGBU on how analytics can have a key impact. SB55 55-Room 108B, CC Airline Revenue Management and Customer Choice Sponsor: Aviation Applications Sponsored Session Chair: Emmanuel Carrier, Delta, emmanuel.carrier@delta.com 1 - Estimation of Airline Itinerary Choice Models using Disaggregate Ticket Data Laurie Garrow, Georgia Institute of Technology, Mason Building, Atlanta, GA, United States of America, laurie.garrow@ce.gatech.edu, Virginie Lurkin, Michael Schyns Airline itinerary choice models support many multi-million dollar decisions, i.e., they are used to evaluate potential route schedules. Classic models suffer from major limitations, most notably they use average fare information but to not correct for price endogeneity. We use a novel database of airline tickets to estimate itinerary choice models using detailed fare data and compare these to classic itinerary choice models that use aggregate fare information but correct for price endogeneity. 2 - Insights from Mining Airline Booking Data Catherine Cleophas, RWTH Aachen, Kackertstrasse 7, 52070, Germany, catherine.cleophas@rwth-aachen.de, Sebastian Vock, Laurie Garrow We present a new perspective on airline booking data by comparing data from geographically and temporally distinct travel itineraries. To this end, we mine several hundred origin-destination-carrier combinations and several thousand itineraries. By clustering booking class distributions, we measure market similarity. The goal is to use cluster-adherence to compute decision trees, so as to investigate the relevance of traditional assumptions about market differentiation. 3 - Nonparametric Estimation of Demand Structures in Airline Revenue Management Johannes Jürg, RWTH Aachen University, Kackertstr. 7, Aachen, 52072, Germany, johannes.ferdinand.joerg@ada.rwth-aachen.de, Catherine Cleophas A central theme of airline revenue management is analyzing historical booking data to draw conclusions on the underlying demand structure. This contribution focuses on the estimation of demand segments present in a market using nonparametrical methods on panel data. We employ finite mixtures to model booking events over time frames and to obtain an estimator for the number of demand segments and their probability distribution over products.

Mihiro Sasaki, Professor, Nanzan University, 18 Yamazato, Showa, Nagoya, 466-8673, Japan, mihiro@nanzan-u.ac.jp, Yoichi Tanaka, Yasuaki Oishi, Masao Fukushima, Masaki Yasunishi

We consider a smart community where the supplier of electricity presents the electricity price varying with different time, and consumers determine their own optimal levels of buying/selling electricity in responding to the time-varying price. We formulate the problem of finding supplier’s optimal pricing as a bilevel programming problem. Computational results show that real-time pricing can effectively control supply and demand of electricity.

SB57 57-Room 109B, CC Electricity Market Models Sponsor: ENRE – Energy I – Electricity Sponsored Session

Chair: Hung Po Chao, Energy Trading Analytics, 2842 Main St., Suite 206, Glastonbury, CT, 06033, United States of America, hungpo.chao@gmail.com 1 - Decentralized Markets with LMP for Efficient Congestion Management of Renewable Electricity Feed-in Hans Schermeyer, Research Associate, Karlsruhe Institute of Technology, Hertzstr. 16, Karlsruhe, 76187, Germany, hans.schermeyer@kit.edu, Valentin Bertsch, Wolf Fichtner In this work, we analyse electricity grid congestion caused by renewables and their necessary curtailment on distribution grid level. To explore the possibilities of enhanced congestion management, we develop an agent-based simulation model that represents a distribution grid in Germany which faces frequent congestion caused by renewables. Inspired by (Distribution) Locational Marginal Pricing theory we implement a decentralized market design for a more efficient congestion management. 2 - Operational Flexibility in Dutch Electricity Markets Robin Broder Hytowitz, Johns Hopkins University, 3400 North Charles Street, Ames 313, Baltimore, MD, United States of America, robin.hytowitz@gmail.com, Ozge Ozdemir, Benjamin Hobbs A two-stage study is performed to evaluate the impact of renewable energy on operational flexibility in the Dutch market given current demand projections. A unit commitment model for a pan-European network simulates the day-ahead market, and the successive balancing market is simulated for the Netherlands alone. The two stages are run for several business cases to analyze resources that can provide further flexibility, including demand-side management, storage, and reserve constraints.

SB56 56-Room 109A, CC

Location Applications Sponsor: Location Analysis Sponsored Session Chair: Mihiro Sasaki, Professor, Nanzan University, 18 Yamazato, Showa, Nagoya, 466-8673, Japan, mihiro@nanzan-u.ac.jp

84