Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

TC35

requests for products, wherein the sizes of products are small enough for consolidation to be an option. To serve these requests, the carrier can solve a variant of the Service Network Design problem wherein each commodity has multiple potential origins. Industrial instances involve a significant number of products, which in turn yields a model that is larger than an off-the-shelf solver can handle. We present a Benders method with a master problem defined over aggregated products. We augment this approach with valid inequalities and heuristics, and with an extensive computational study illustrating its effectiveness. n TC33 North Bldg 222C Big Data and Machine Learning for Advanced Manufacturing Emerging Topic: OR and Advanced Manufacturing Emerging Topic Session Chair: Guha Manogharan, Pennsylvania State University, University Park, PA, United States Co-Chair: Michael G. Kay, North Carolina State University, Raleigh, NC, 27695, United States 1 - Uncertainty Quantification for Integrated Simulation Models with Unobservable Variables Using Bayesian Networks Mohamad Mahmoudi, Texas A&M University, 1221 April Bloom, A, College Station, TX, 77840, United States, Alaa Elwany In this work we investigate the problem of uncertainty quantification for a system of computer models which are integrated in a hierarchical fashion. We consider the assumption of existence of unobservable outputs in our system. The proposed methodology employs Bayesian networks to conduct uncertainty quantification. To evaluate the performance of the proposed method, a case study of simulation models used in metal additive manufacturing is presented where two high-fidelity models are integrated and uncertainty sources for each model is identified. The results of using the Bayesian network for uncertainty quantification is the presented and thoroughly discussed. 2 - Revitalizing Sand-Casting Supply Chain in the United States: A Study on Locating 3D-Sand Printer Hubs to Support Metal Foundries Michael Kay, Pennsylvania State University, 232 Reber Building, State College, PA, 16802, United States, Ramin Ahmed, Casey Bates, Paul Lynch, Guha Prasanna Manogharan Additive manufacturing (AM), over the past few years, has played a revolutionary role in the field of manufacturing, mainly in the casting process. In casting industries, sand casting provides huge scope due to its expandability and ability to create complex metal parts of almost any alloy. Recent research has determined various benefits of 3D printing sand molds including time efficiency, minimizing product and development delay, the ability to form very intricate castings, reduction in the need for inventory and so on. While all these benefits are apparent, traditional manufacturers are still not encouraged to adopt only due to its high fixed cost and lack of expertise. On the other hand, there are so much demand over the U.S it is hard only a handful of dedicated AM firms to the demand points. This is due to the fact that, the transportation cost will become prohibitive due to the fragile nature of the sand molds. In this paper, we investigate a system of strategically-located AM hubs which can encompass all the demand points. Using North American Industry Classification System (NAICS) data for the foundries in the U.S., we have used an uncapacitated facility location model to determine the optimal locations for AM hubs where we considered the geographic data and demand and strategically identified the fixed cost using various data like AM machine cost, packaging cost, employee salaries, rent and other consumables which were obtained through personal communication. Results from this study will identify: (a) candidate US counties to build AM hubs, (b) total cost (fixed, operational and transportation). We also noticed that over the past years the demand has been increasing, which would mean an increased capacity to satisfy those demand. We hope to address this issue with a sensitivity analysis in 3 levels: (low, medium, high). With each such level of increase, we aim to demonstrate how and where the additional hubs should be added. n TC34 North Bldg 223 12:05 - 12:50 Optimization Firm/ 12:50 - 1:35 MathWorks Emerging Topic: Technology Tutorials Emerging Topic Session 1 - Experiential Machine Learning with MATLAB Elvira Osuna-Highley, Mathworks, Natick, MA, United States There is a growing need to teach the next generation of scientists and engineers ever more complex concepts. It is critical that intuition is built quickly so that

they can apply their knowledge to develop the technology of the future. This calls for experiential learning techniques that make concepts easy and fun to teach, learn and test. This tutorial will illustrate experiential learning using the Classification Learner app in MATLAB. You will see how to work through modules in MATLAB in order to build various classification models. This will give you the opportunity to understand the potential for using an experiential approach in the classroom as well as introducing you to tools that you may use in your research. 2 - MathWorks – Information not available.

n TC35 North Bldg 224A Urban Air Mobility Sponsored: Aviation Applications Sponsored Session

Chair: Peng Wei, Iowa State University, Ames, IA, 50011, United States 1 - Energy Optimal Speed Profile for Arrival of Tandem Tilt-Wing eVTOL Aircraft with RTA Constraint Priyank Pradeep, Iowa State University, Ames, IA, 50011, United States, Peng Wei The electric vertical takeoff and landing (eVTOL) aircraft can alleviate transportation congestion on the ground by utilizing three-dimensional airspace efficiently. However, the endurance of Lithium-ion Polymer (Li-Po) batteries imposes severe constraints on the operational time span of an eVTOL on urban air mobility (UAM) passenger transportation mission. This research focuses on the problem formulation and numerical solution of a fixed final time multiphase optimal control problem with energy consumption as the performance index for a tandem tilt-wing eVTOL. 2 - Adaptive Arrival Sequencing and Scheduling for Mixed Fleet eVTOL Aircraft in On-Demand Urban Air Mobility Priyank Pradeep, Iowa State University, Ames, IA, 50011, United States Our current research is focussed on solving Aircraft Sequencing and Scheduling (ASS) problem for eVTOL aircraft in Urban Air Mobility (UAM) environment. The concept of operations (CONOPs) for both types of eVTOL aircraft (wingless and winged) is assumed to be cruising at constant altitude followed by the vertical descent. To compute the set of RTAs, first, the sequencing order of landing of the eVTOL aircraft will be calculated using heuristic methods, such as Genetic Algorithm (GA). Once we determine an arrival sequencing order, the arrival scheduling sub-problem can be formulated as mixed integer linear programming (MILP) problem and will be solved using open source MILP solver Gurobi Optimizer. 3 - Autonomous On-Demand Free Flight Operations in Urban Air Mobility using Monte Carlo Tree Search Xuxi Yang, Iowa State University, Ames, IA, 50011, United States, Peng Wei 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 - Innovating Airline Operations Centers to Support New Fleet Demands Victoria C. Nneji, Duke University, Durham, NC, 27708, United States Several manufacturers are developing new vehicles for high-speed intra-city air taxi on-demand mobility (ODM) services. With diverse vehicle design requirements, how might airline operations centers need to innovate to support these new fleet demands? We used task- and time-based data from a collective case study of dispatchers to develop a discrete event simulation to serve as a predictive model of human-system performance in these air operations centers. With this tool, companies can rapidly prototype future ODM concepts of operations to make better informed strategic, tactical, and operational decisions in staffing and designing centers. 5 - A Progressive Trajectory Planning Algorithm for High-Density 2D Traffic Yanchao Liu, Wayne State University, Detroit, MI, United States This talk presents a trajectory planning algorithm based on nonlinear optimization techniques. The model centrally coordinates trajectories for all vehicles traversing a shared two-dimensional space. To avoid path deadlocks associated with local optima, a progressive solution algorithm is developed. In addition, a set of new metrics is proposed to measure traffic flow efficiency of 2D transit systems. Simulation results will be demonstrated.

319