Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SD70

2 - The Dynamic Impact of Quantity Restriciton on Backers’ Investment Intention Zhijin Zhou, University of Washington, Foster School of Business, Seattle, WA, 98195, United States, Chaoliang Ma, Yong Tan We analyze the effect of implementing quantity restriction, a prevalent marketing strategy to promote sales in the off-line market, in the context of crowdfunding. Using a dynamic panel dataset, we start with a preliminary analysis to get insights into backers’ basic response patterns. Then, we extend our model and allows for (a) varying parameters to fully capture the underlying dynamics, (b) correlation in the errors that might affect parameter evolving process. Our results suggest that the scarcity of a product increases backers’ overall evaluation of the product, while it plays a moderating role in attenuating backers’ reliance on peers’ action and decreasing their price sensitivity. 3 - A Model of Smart Technologies Xinxin Li, University of Connecticut, 2100 Hillside Road U-1041, Storrs, CT, 06269, United States, Yuxin Chen, Monic Sun We study the optimal pricing and design of smart technologies that are based on artificial intelligence (AI) and can learn consumers’ preferences over time. Our preliminary analysis suggests that it is not always profitable to increase the smartness of a firm’s technology even when doing so does not involve direct costs. The “price in our model can be interpreted as either a direct price that consumers have to pay to the firm or a form of advertising exposure. Correspondingly, our model has implications not only for the pricing and design of smart technologies and their interaction with consumers, but also for platforms on which advertisers aim to target the users of smart technologies. 4 - Too Much of a Good Thing? Wei Zhou, PhD Student, University of Arizona, Tucson, AZ, 85721, United States, Mingfeng Lin The importance of reputation online is well established and is especially true for service-oriented markets spanning such as professional services. We explore limits to such claims using a detailed transactional dataset and discuss implications for designing such reputation systems. Chair: Yu Ding, Texas A&M University, ETB 4016, MS 3131, Texas A&M University, College Station, TX, 77843-3131, United States 1 - Causation-based Process Monitoring and Diagnosis for Multivariate Categorical Processes Jian Li, School of Management, Xi’an Jiaotong University, No 28 Xianning West Road, Xi’an, China Many applications involve causal relationships among multiple categorical variables/factors, where shifts at cause factors will propagate to their effect factors. A causation-based rather than correlation-based description would better account for such causal relationships. We integrate a Bayesian network and construct one general and one directional control charts for detecting shifts in the conditional probabilities of categorical factors. Simulations have demonstrated their effectiveness. 2 - Ensemble Bayesian SPC: Multi-Mode Process Monitoring for Novelty Detection Irad Ben-Gal, Tel Aviv University, Tel-Aviv, Israel, Marcelo Bacher. We propose a monitoring method based on a Bayesian analysis of an ensemble- of-classifiers for Statistical Process Control (SPC) of multi-mode systems. A specific case is considered, in which new modes of operations (new classes), also called “novelties,” are identified during the monitoring stage of the system. The proposed Ensemble-Bayesian SPC (EB-SPC) models the known operating modes by categorizing their corresponding observations into data classes that are detected during the training stage. Ensembles of decision trees are trained over replicated subspaces of features, with class-dependent thresholds being computed and used to detect novelties. In contrast with existing monitoring approaches that often focus on a single operating mode as the “in-control” class, the EB-SPC exploits the joint information of the trained classes and combines the posterior probabilities of various classifiers by using a “mixture-of-experts” approach. Performance evaluation on real datasets from both public repositories and real- world semiconductor datasets shows that the EB-SPC outperforms both conventional multivariate SPC as well as ensemble-of-classifiers methods and has a high potential for novelty detection including the monitoring of multimode systems. 3 - Discussant Matthew Plumlee, Northwestern University, IL, United States This talk will discuss the two other papers in this session, discussing acheivments, potential drawbacks, connections to other literature, and possible extensions. n SD68 West Bldg 105C IISE Transactions Invited Session Sponsored: Quality, Statistics and Reliability Sponsored Session

n SD69 West Bldg 106A Functional Data or Profiled Response Analysis Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Rong Pan, Arizona State University, Tempe, AZ, 85287-8809, United States 1 - An Efficient Surrogate Model for Emulation and Physics Extraction of Large Eddy Simulations Chih-Li Sung, PhD, Georgia Institute of Technology, Atlanta, GA, United States, Simon Mak, Xingjian Wang, Shiang-Ting Yeh, Yu- Hung Chang, Roshan J. Vengazhiyil, Vigor Yang, C. F. Jeff Wu In the quest for advanced propulsion and power-generation systems, high-fidelity simulations are too computationally expensive to survey the desired design space. In this paper, we propose a new surrogate model that provides efficient prediction and uncertainty quantification of turbulent flows in swirl injectors with varying geometries. The novelty of the proposed method lies in the incorporation of known physical properties of the fluid flow as simplifying assumptions for the statistical model. In view of the massive simulation data at hand, which is on the order of hundreds of gigabytes, these assumptions allow for accurate flow predictions in around an hour of computation time. 2 - Dynamic Curve Alignment Based on Penalized-spline Smoothing Kaibo Wang, Professor, Tsinghua University, Department of Industrial Engineering, Tsinghua University, Beijing, 100084, China Unaligned profiles with amplitude and phase variabilities have to be registered (aligned) through shifting, time warping or coordinate alignment so that samples are comparable and easy to handle. This work proposes a penalized-spline smoothing method for profile alignment. The strategy is try to capture the smoothness and spatially correlated features of warping shifts through a penalized registration function. A dynamic programming algorithm is developed to obtain the optimal path. 3 - Nearest-neighbor Gaussian Process Emulation for Tensor Responses in Freeze Nano Printing Hongyue Sun, University at Buffalo, 319 Bell Hall, Industrial and Systems Engineering, Buffalo, NY, 14260, United States, Guanglei Zhao, Chi Zhou Existing energy storage devices have either high power density or high energy density, but not both. Freeze nano printing can potentially solve the problem by combining inkjet printing and freeze casting to print multi-scale porous structures. In this process, the thermal history is fundamental to its quality and productivity. The thermal history evolves over space and time, and cannot be fully captured by existing sensors. The physical simulation models can describe the spatial-temporal thermal history (i.e., tensor response) but are computationally demanding. We propose a Nearest-Neighbor Gaussian Process (NNGP) based emulator to address the computation problem for the tensor response. 4 - Experimental Designs for Studying Dynamic Response Modeled by B-splines Rong Pan, Arizona State University, School of Computing Informatics & Decison Sys, P.O. Box 878809, Tempe, AZ, 85287- 8809, United States In this talk we discuss the B-spline models used for modeling dynamic responses from industrial experiments and how to design the experiment to collect system responses more efficiently.

n SD70 West Bldg 106B

Joint Session QSR/DM: Machine Learning Based Approaches for Semiconductor Manufacturing Processes Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Youngseon Jeong, Chonnam National University, 5012 King David Boulvard, Annandale, VA, 22003-4033, United States Co-Chair: Jeongsub Choi, Rutgers-State University of New Jersey, Highland Park, NJ, 08904, United States

115