Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MA35

2 - Dynamic Design of Reserve Crew Pairings for Long Haul Airline Crew Lennart Scherp, PhD Candidate, Delft University ot Technology, Delft, Netherlands, Richard Janssen, Bruno F. Santos In this work we propose a new reserve crew pairings generation algorithm. A method is used that iteratively selects a set of pairings and evaluates them based on simulations of roster disruptions to obtain the most efficient set of reserve pairings. The goal is to minimize the effects of disruptions on rosters. Results show improvements of about 5% over the current approach used by a large European airline. 3 - Airline Crew Scheduling with Re-timing and Complex Aircraft Connection Rules Fredrik Altenstedt, Optimization Expert, Jeppesen, Odinsgatan 9, G÷teborg, SE-411 03, Sweden This presentation considers finding crew trips covering all flights for an airline given that we may adjust the flight times slightly, commonly known as the pairing problem with retiming. The aircraft routes influence the crew rules, shorter connections are allowed if the crew does not change aircraft. Both re- timings and short connections may require extra aircraft. This is normally prevented using plane count constraints, but these only work for simple connection rules. We give an overview of how we maintain aircraft feasibility for more complex rules by generating aircraft consistency constraints dynamically as well as share our experience from using the system with one of our clients. 4 - Shift Scheduling with Flexible Execution Times and Sequence Dependent Travel Pouya Barahimi, Oregon State University, Corvallis, OR, 97331, United States, Hector Vergara We consider the scheduling problem where each task may require multiple workers with different skills. Tasks have fixed duration and can be executed within a given time window. Travel times between tasks are different. The objective is to maximize the reward gained by executing tasks given a limited team of workers. A branch-and-price algorithm and two heuristics are developed to solve instances. A numerical study assessing the effect of inputs on run-time and quality of the solutions obtained is presented. The problem is presented in the context of the airline industry where ground crew located at an airport perform tasks to get flights ready for departure. 5 - Modeling Flight Effort in Air Operations Training Programs Sergio Rebou as, Technological Institute of Aeronautics, Pra a Marechal Eduardo Gomes, 50, Sao Jose dos Campos, Brazil, Talita Alessandra da Silva, Fernando Teixeira Abrah π o The flight effort forecasting in pilot training programs is a challenge for civil and military organizations. The mobilization of material and human resources to support air operations is the main factor of impact in their economic health and administrative efforts. The logistic and operational support contracts are based on this forecast, whose inaccuracy would imply significant financial and administrative consequences. This article describes a method to estimate flight effort based on a pilot’s training program, integrating the individual instructions, operational and maintenance constraints and historical data, with the objective of maximizing its accuracy. n MA37 North Bldg 225A Modern Approaches to Stochastic Control Sponsored: Applied Probability Sponsored Session Chair: Vivek Farias, Massachusetts Institute of Technology, Cambridge, MA, 02139, United States Co-Chair: Ciamac Cyrus Moallemi, Columbia University, New York, NY, 10027, United States 1 - Reinforcement Learning in Portfolio Optimization Eli Gutin, Massachusetts Institute of Technology, Cambridge, MA, 02139, United States, Vivek Farias, Ciamac Cyrus Moallemi In this talk we propose a general deep RL, policy-gradient based algorithm for tackling high-dimensional control problems with continuous state and action spaces, and quasi-linear dynamics. We demonstrate state-of-the-art performance on key benchmark problems in dynamic portfolio optimization and algorithmic trading. On some instances we see a provable near-zero optimality gap, compared to existing cutting-edge approaches that achieve around a 5% gap.

n MA35 North Bldg 224A AAS Best Student Presentation Competition I Sponsored: Aviation Applications Sponsored Session Chair: Susan Hotle, Virginia Polytechnic Institute and State University. 1 - Impact of Payload Amount on Battery Consumption Rate in a Delivery Application of Drones Maryam Torabbeigi, University of Houston, Houston, TX, United States The drone battery charge limitation is an important factor in drone scheduling in order not to run out of battery during the flight. This study investigates the relationship between battery consumption rate (BCR) and the payload amount, and also the impact of payload amount (customer’s demand) on the drone scheduling. The collected data verifies a linear relationship between BCR and the payload amount. A routing problem is proposed for the drone scheduling. The model determines the number of drones, their path, the assigned customers, and the battery charge at each flight segment. The results show the impact of including BCR in the scheduling. 2 - Airline Passenger Route Share Forecast Xufang Zheng, Iowa State University, 537 Bissell Road #2362, Ames, IA, 50010, United States Airline passenger route share(directShare) is the ratio of direct passengers to total passengers on O&D level. It is an important feature of passenger flow distribution. DirectShare is an O&D specific feature, which is highly correlated with quarterly lag. Various supervised learning methods are carefully explored. The best model is gradient boosting machine(GBM), which has better prediction performance than FAA TAF-M directShareforecast model. Category based learning is newly proposed, which provides better prediction performance than GBM. The C- basedapTC model is the best category based learning model, which can provide a long term directShare forecast with less fluctuations. 3 - Autonomous On-Demand Free Flight Operations in Urban Air Mobility Using Monte Carlo Tree Search Xuxi Yang, Iowa State University, Howe Hall, 537 Bissell Rd, Ames, IA, 50011, United States Vertical takeoff and landing (VTOL) aircraft for on-demand air taxi will bring fundamental changes to daily commutes. NASA, Uber, and Airbus have been exploring the exciting concept of Urban Air Mobility (UAM). In order to enable safe and efficient autonomous on-demand free flight operations in this UAM concept, a computational guidance algorithm was designed and analyzed with collision avoidance capability. The approach is to formulate this problem as a Markov Decision Process and solve it using Monte Carlo Tree Search. A simplified numerical experiment was created and results show that this algorithm can help aircraft quickly reach the trip destination and avoid conflicts with other aircraft. 4 - Safe and Efficient Arrival of EVTOLS in On-Demand UAM Priyank Pradeep, Iowa State University, 537 Bissell Road, Howe Hall, Ames, IA, 50011, United States The electric vertical takeoff and landing (eVTOL) aircraft can alleviate ground congestion by utilizing three-dimensional airspace. However, the endurance of Lithium-ion Polymer (Li-Po) batteries imposes constraints on the operational time span. The first part focuses on formulations of a fixed final time multiphase optimal control problem with energy consumption for the eVTOL multirotor and tandem tilt-wing aircraft types. The second part involves a sequencing problem for a mixed fleet of eVTOLs scheduled to land on a vertiport with single landing slot. A sensitivity analysis is performed to see the impact of the fleet mixture and resequencing on the landing completion time. Airline Crew Management Sponsored: Aviation Applications Sponsored Session Chair: Bruno F. Santos, TU Delft, Delft, 2629 HS, Netherlands 1 - Generating Crew Pairings for Cargo Aviation Junhong Guo, University of Michigan, Ann Arbor, M I, United States, Amy Cohn We consider a variation of the airline crew pairing problem in which we focus on the challenges faced by a cargo carrier with predominantly long-haul international flights. The difference in rules and pairing characteristics relative to domestic passenger aviation (e.g. pairings are typically on the order of 15 days) provide both challenges and opportunities. We present heuristics and optimization-based approaches for improving this process. n MA36 North Bldg 224B Joint Session AAS/Practice Curated:

132