Informs Annual Meeting 2017

WA67

INFORMS Houston – 2017

3 - Multiple Change-point Modeling and Exact Bayesian Inference of Degradation Signal for Prognostics Improvement Yuxin Wen, PhD Student, University of Texas-El Paso, El Paso, TX, 79968, United States, ywen@miners.utep.edu We propose a Bayesian multiple change-point modeling framework to better capture the degradation path for prognostics improvement. At the offline modeling stage, a novel stochastic process is proposed to model the joint prior of change-points and positions. At the online monitoring and RUL prediction stage, a recursive updating algorithm is developed to exactly calculate the posterior distribution and RUL prediction sequentially. To control the computational cost, a fixed-support-size strategy at the online stage and a partial Monte Carlo strategy in the RUL prediction are proposed. The effectiveness and advantages of the proposed method are demonstrated simulation and real case studies. 4 - Economic Impact of Raised Medians on Businesses in South Carolina Samaneh Shiri, University of South Carolina, 1076 Pavilion Tower Circle, Columbia, SC, 29201-2361, United States, shiri@email.sc.edu, Mohammad Torkjazi, Nathan Huynh Access management techniques affect economic activities along highway corridors. Raised medians that restrict left turns at mid-block are often considered as an effective access management technique which has been shown to improve traffic operations and safety. However, such restrictions also limit access to driveways. In this study, the economic impact of raised medians on businesses in South Carolina is examined using two methods: (1) surveys to study the perception of businesses and customers about raised medians, and (2) an assessment of the actual impact of a raised median on sales volume of adjacent businesses. 371B Data Analytics for System Improvement III Sponsored: Quality, Statistics and Reliability Sponsored Session Chair: Xi Zhang, Peking University, Beijing, China, xi.zhang@pku.edu.cn Co-Chair: Kaibo Liu, UW-Madison, Madison, WI, 53706, United States, jacoblkb@gmail.com 1 - Stochastic Modeling and Diagnosis of Leak Areas for Surface Assembly Jie Ren, Graduate Research Assistant, Florida State Unigersity, Tallahassee, FL, 32310, United States, jr14r@my.fsu.edu, Hui Wang, Chiwoo Park Assembly through mating a pair of machined surfaces plays a crucial role in automotive powertrain production, and the mating errors during the assembly can cause internal leakage and functional problems. The surface mating errors are difficult to diagnose because they are not measurable. To address these limitations, this talk presents a novel diagnostic method to predict potential leak areas given the measurements on the profiles of mating surfaces. The effectiveness of the proposed method is verified by case studies. The approach provides practical guidance for the subsequent assembly process as well as troubleshooting in surface machining processes. 2 - Causation-based Monitoring and Diagnosis for Multivariate Categorical Processes with Ordinal Information Xiaochen Xian, UW-Madison, Madison, WI, United States, xxian@wisc.edu, Jian Li, Kaibo Liu The monitoring and diagnosis of multivariate categorical processes (MCPs) have drawn increasing attention lately. In these applications, there usually exists natural order among the attribute levels of some categorical variables, such as good, neutral, and bad for measuring the product quality. In this paper, we leverage Bayesian networks to characterize MCPs with a causal structure, where the categorical variables can be either nominal, ordinal, or a combination of both. We develop one general control chart and one directional control chart, both of which fully exploit the causal relationships and the ordinal information for better process monitoring and diagnosis. 3 - System Redesign under Unknown Demand Rates Qianru Ge, TU/e, Paviljoen N.5, IE&IS, Den Dolech 2, Postbus 5, Eindhoven, 5612 AZ, Netherlands, q.ge@tue.nl, Willem van Jaarsveld, Zumbul Atan We consider a newly developed system sold under a PBC. The system is subject to failures with a random failure rate. The OEM is responsible for the availability of the system and makes a redesign decision at each decision period. The trade-off is that low failure rates require high redesign costs while high failure rates result in high repair and penalty costs. We use a MDP model to support the decision. As more information become available over time, we use Bayesian model to update the failure rate distribution. We show that the optimal redesign policy has a nondecreasing control limit structure. We benchmark the optimal policy against the policy only optimizing the initial design but with no redesigns. WA67

4 - Estimation of Spatiotemporally-varying Parameters in Partial Differential Equation Models Xi Zhang, Peking University, 298 Chengfu Road, Founder Building 512, Beijing, China, xi.zhang@pku.edu.cn, Di Wang Estimation of temperature field provides a prerequisite opportunity to understand the dynamics of complex systems in many areas. By combining the data collected from sensor networks, the accuracy of the estimation could be greatly improved. This article presents an estimation method to model the spatiotemporal dynamics of a temperature field by integrating a physical dynamics model and a spatiotemporal Gaussian process. Results of the studies give some insight into the spatiotemporal dynamics of a temperature field and provide guidance to the optimum design of engineering structures to conserve energy and reduce production cost.

WA68

371C Quality, Statistics and Reliability Sponsored: Quality, Statistics and Reliability Sponsored Session

Chair: Anahita Khojandi, Assistant Professor, University of Tennessee, Knoxville, TN, 37996, United States, anahitakhojandi@gmail.com 1 - Optimizing Condition based Maintenance for Systems with Degrading Sensors Anahita Khojandi, University of Tennessee, 521 Tickle Building, Knoxville, TN, 37996, United States, khojandi@utk.edu, Mahboubeh Madadi We consider a degrading system with non-silent failures. System failure is costly and it requires an immediate system replacement. The system status may be partially observed using a set of heterogeneous sensors at a given cost. The sensors are noisy and their noise level is a function of the sensor status. Our goal is to develop a POMDP model to determine the optimal condition based maintenance scheme to minimize the total expected discounted cost. 2 - A Decomposition Algorithm to Measure Redundancy in Structured Linear Systems Vishnu Vijayaraghavan, Texas A&M.University, 400 Nagle Street, College Station, TX, 77840, United States, vishnunitr@tamu.edu, Kiavash Kianfar, Yu Ding, Hamid R. Parsaei The degree of redundancy of linear sensor systems is a measure of robustness of the system against sensor failures. Finding the degree of redundancy for structured linear systems is proven to be NP-hard. Bound and decompose, mixed integer programming, l1-minimization methods have all been studied and compared in the literature. We propose a mixed integer programming based decomposition technique that utilizes the bordered block diagonal structure inherent in most practical linear sensor systems to decompose the problem into much smaller sub-problems on which the computations are performed, thereby leading to significant reductions in run time. 3 - Hierarchical Spatially Varying Coefficient Process Model Heeyoung Kim, Assistant Professor, Industrial and Systems Engineering, Daejon, Korea, Republic of, heeyoungkim@kaist.ac.kr The spatially varying coefficient process model is a non-stationary approach to explaining spatial heterogeneity by allowing coefficients to vary across space. In this study, we develop a methodology for generalizing this model to accommodate geographically hierarchical data. We consider two-level hierarchical structures and allow for the coefficients of both low-level and high-level units to vary over space. 4 - Optimizing Periodic Inspection Frequencies for a Class of Stochastically Degrading Systems Mahboubeh Madadi, Louisiana Tech University, College of Engineering and Science, P.O. Box 10348, Ruston, LA, 71272, United States, madadi@latech.edu, David Kaufman, Murat Kurt We consider existing models that optimize repair-replacement decisions for systems the degradation status of which follow a discrete time Markov chain over a set of finite states and can be revealed only by costly inspections. Given worse conditions imply higher operation costs, we utilize first-order stochastic dominance relationship among the powers of IFR-structured degradation matrices to propose approximately-optimal periodic inspection decisions that minimize the total expected discounted cost due to the operation, repair, and inspection. We illustrate our approach through numerical examples.

466