2016 INFORMS Annual Meeting Program

WA62

INFORMS Nashville – 2016

2 - Schedule Flexibility And Shared Corridor Capacity Darkhan Mussanov, University of Illinois at Urbana - Champaign, Urbana, IL, 61801, mussano2@illinois.edu North American railways operate in the unstructured manner, i.e without strictly adhered to timetable and pre-planned meets. Unscheduled operation poses a challenge the railway planners by increasing the number of possible meets on railway track. This research aimed to investigate the relationship between the schedule flexibility, infrastructure investment and level of service. The results revealed the fragile nature of the structured operation and the substantial investment requirements for the small shifts toward unscheduled operation. 3 - Siding Length, Train Length, And Capacity For Additional Traffic On Single Track Lines Bradford Kippen, University of Illinois at Urbana - Champaign, Urbana, IL, 61801, United States, kippen2@illinois.edu Operation of longer freight trains on single track lines has allowed North American Class 1 Railroads to expand freight car throughput without investment in additional infrastructure. However, implementation of this strategy is often limited when the length of longer trains exceeds length of existing passing sidings. RTC analysis has been used to model train delay patterns associated with various strategies of implementing long trains on a hypothetical corridor. In addition, following a simulated increase in traffic volume, a model was developed to determine the infrastructure investment required in either new or longer sidings to return to a baseline level of service. 4 - Capacity Allocation In Vertically Integrated Rail Systems: A Bargaining Approach This paper presents a game-theoretic bargaining approach to allocating rail line capacity in vertically integrated systems. A passenger rail agency (PRA) negotiates with the host freight railroad (FRR) to determine train schedules and the associated payment. Bargaining in both complete and incomplete information settings are considered; the latter arises because FRR may withhold its private cost information. Equilibrium schedule with complete information, maximizes system welfare. With incomplete information, PRA may choose between pooling and separating equilibrium strategies while proposing a payment, depending on its prior belief the cost type of FRR. WA62 Cumberland 4- Omni Air Traffic Management and Airline Operations Sponsored: Aviation Applications Sponsored Session Chair: Peng Wei, Iowa State University, 2312 Howe Hall, 537 Bissell Road, Ames, IA, 50011, United States, pwei@iastate.edu 1 - Terminal Area Sequencing And Scheduling: The Single Runway Case Jitamitra Desai, Professor, Nanyang Technological University, Singapore, Singapore, jdesai@ntu.edu.sg, Rakesh Prakash This paper addresses the aircraft sequencing problem over the entire terminal maneuvering area (TMA) under a mixed-mode, single runway operating scenario. In contrast with existing approaches that only consider the runway as a bottleneck, our 0-1 mixed-integer LP formulation optimizes flight sequences and schedules by taking into account the configuration and associated constraints of the entire TMA region. Variable fixing strategies and valid inequalities are derived to tighten the continuous relaxation of the problem. Computational results show the overall delay in the system can be reduced by nearly a 30% margin over the default FCFS policy and by nearly 10% over the runway sequencing policy. 2 - Passenger Route Choice Prediction In The U.S. Airline Industry: Statistical Methods versus Machine Learning Techniques Chia-Mei Liu, FAA, Chia-Mei.Liu@faa.gov, Peng Wei, Jerrod Sharpe In the community of airline forecast research, while there are plenty research in airline passenger demand modeling that designs to forecast passenger growth, relatively little attention has been paid to passenger route choice forecast between nonstop and connecting flights. This paper contributes to this area of the research by constructing a route choice model, estimated through statistical methods as well as machine learning methods. The results have important implications in operation forecast as route choice forecast affects airline fleet planning. Consequently, this research benefits policy makers and industry practitioners by expanding our understanding on passenger route choice. Bo Zou, University of Illinois at Chicago, Chicago, IL, United States, bzou@uic.edu, Ahmadreza Talebian

3 - Capturing Passenger Compensation Impacts For An Integrated Airline Recovery Luis Cadarso, Rey Juan Carlos University, Camino del Molino s/n, Fuenlabrada, 28943, Spain, luis.cadarso@urjc.es, Vikrant Vaze The European flight delay compensation regulation (EC) No 261/2004 establishes common rules on compensation to passengers in the event of disruptions. We develop an integrated approach that recovers airline timetable, fleet assignment, aircraft routings, and passenger itineraries capturing the impacts of airlines’ decisions on passenger compensation. We evaluate scenarios involving disruptions, and optimize recovery decisions to maximize profits by modeling passenger no-shows after disruptions.

WA63 Cumberland 5- Omni Location Models Sponsored: Location Analysis Sponsored Session

Chair: Oded Berman, University of Toronto, 105 St. George Street, Toronto, ON, M5S 3E6, Canada, berman@rotman.utoronto.ca 1 - Responsive Supply Chain Network Design Oded Berman, University of Toronto, 105 Saint George Street, Toronto, ON, M5S 3E6, Canada, berman@rotman.utoronto.ca, Robert Aboolian, Jiamin Wang In this paper, we address the network design of a responsive supply chain consisting of make-to-order facilities facing stochastic demand. Each facility has a finite capacity and stochasticity of demand may lead to congestion delays at the facilities. We consider three problems. In the first, we minimize the total network cost including delivery and capacity costs while maintaining an acceptable response time to customers. In the second, a penalty is charged on the number of units delivered later than the targeted response time. In the third, the penalty charged also depends on the number of days that the delivery is late. The penalty cost in both problems 2 and 3 are a function of network’s response time. 2 - The p-center On A Network With Probabilistic Demand Weights We study the p-center problem on a network with probabilistic demand weights. Two models are presented. The objective of the first model is to maximize the probability that the longest weighted distance from the nodes to the closest facility does not exceed a pre-selected target level. In the second model, facilities are located so as to minimize the value-at-risk, namely, a quantile of the longest weighted distance with a specified confidence level. Special cases are identified that are equivalent to the deterministic center model. The problem is shown to be NP-hard. Exact solution procedures and heuristics are developed for demand weights of discrete and continuous probability distributions. 3 - Location Depots To Facilitate Routing A Mixed Fleet Of Electric And Conventional Vehicles Nan Ding, Univesity of Buffalo (SUNY), Buffalo, NY, nanding@buffalo.edu, Rajan Batta Most of current works of routing electric vehicles (EVs), assuming charging availability en route, may need high cost of establishing charging infrastructures. In this work, an alternative strategy to adopt EVs is proposed. This strategy considers charging to be only allowed at depots over the night. To this end, intermediate depots (IDs) are introduced to facilitate routing EVs to customers while conventional vehicles are used to serve IDs from center depot. To determine optimal locations of IDs and routing plans, a joint location-routing problem is formulated. A bi-level heuristic method with upper level determining the locations of IDs and lower level determining routing plans is developed. Jiamin Wang, Long Island University, Brookville, NY, United States, Jiamin.Wang@liu.edu, Oded Berman

384