Abe Zeid

Solving large scale disassembly line balancing problem with uncertainty using reinforcement learning

Due to increasing environmental concerns, manufacturers are forced to take back their products at the end of products’ useful functional life. Manufacturers explore various options including disassembly operations to recover components and subassemblies for reuse, remanufacture, and recycle to extend the life of materials in use and cut down the disposal volume. However, disassembly operations are problematic due to high degree of uncertainty associated with the quality and configuration of product returns. In this research we address the disassembly line balancing problem (DLBP) using a Monte-Carlo based reinforcement learning technique. This reinforcement learning approach is tailored fit to the underlying dynamics of a DLBP. The research results indicate that the reinforcement learning based method is able to perform effectively, even on a complex large scale problem, within a reasonable amount of computational time. The proposed method performed on par or better than the benchmark methods for solving DLBP reported in the literature. Unlike other methods which are usually limited deterministic environments, the reinforcement learning based method is able to operate in deterministic as well as stochastic environments.

Mapping Product Design Specification for Mass Customization

Mass customization is an emerging field in manufacturing research community where customer satisfaction is achieved by complying with exact customer requirements at mass production efficiencies. Product varieties have mostly failed in the aspect of satisfying customer requirements, since knowing beforehand exactly what customer needs is beyond the scope of any manufacturing or marketing paradigm. So design to order and build to order is widely talked in the industry to capture heterogeneous market segments. In order to achieve the challenging task of customization at mass production efficiencies, a tight integration of the various entities involved is necessary. Gathering customer requirements, finalizing product design and eventual manufacturing need to be seamlessly integrated to achieve mass production efficiencies. Product designs are classified into customizable product platform families, and then searched based on customer requirements to identify the most conformal design family. Finally the design parameter transport is carried out from customer domain to product domain. Eventually a valid and realizable product specification is generated. The paper will address these issues of classification, selection and ultimately the mapping of parameters. A methodology to classify the product design information, which can easily accommodate design variations based on product platform architecture is proposed. Further, adaptive design customization, which relates most design parameters with the scalable platform design parameters using matrix formulation, is discussed. The proposed design parameter classification and adaptive synthesis of design parameters is applied to a spring design example.

The structure and analysis of nanotechnology co-author and citation networks

Research activities and collaborations in nanoscale science and engineering have major implications for advancing technological frontiers in many fields including medicine, electronics, energy, and communication. The National Nanotechnology Initiative (NNI) promotes efforts to cultivate effective research and collaborations among nano scientists and engineers to accelerate the advancement of nanotechnology and its commercialization. As of August 2008, there have been over 800 products considered to benefit from nanotechnology directly or indirectly. However, today’s accomplishments in nanotechnology cannot be transformed into commercial products without productive collaborations among experts from disparate research areas such as chemistry, physics, math, biology, engineering, manufacturing, environmental sciences, and social sciences. To study the patterns of collaboration, we build and analyze the collaboration network of scientists and engineers who conduct research in nanotechnology. We study the structure of information flow through citation network of papers authored by nano area scientists. We believe that the study of nano area co-author and paper citation networks improve our understanding of patterns and trends of the current research efforts in this field. We construct these networks based on the publication data collected for years ranging 1993 through 2008 from the scientific literature database “Web of Science”. We explore those networks to find out whether they follow power-law degree distributions and/or if they have a signature of hierarchy. We investigate the small-world characteristics and the existence of possible community structures in those networks. We estimate the statistical properties of the networks and interpret their significance with respect to the nano field.

Nanomanufacturing systems: opportunities for industrial engineers

SATISH BUKKAPATNAM1,∗, SAGAR KAMARTHI2,∗, QIANG HUANG3,∗, ABE ZEID2 and RANGA KOMANDURI4 1School of Industrial Engineering and Management, Oklahoma State University, Stillwater, OK 74078, USA E-mail: bsatish@okstate.edu 2Department of Mechanical and Industrial Engineering, Northeastern University, Boston, MA 02115, USA 3Department of Industrial and Systems Engineering, University of Southern California, Los Angeles, CA 90089, USA 4School of Mechanical and Aerospace Engineering, Oklahoma State University, Stillwater, OK 74078, USA

Analysis of Molecular Beam Epitaxy Process for Growing Nanoscale Magnesium Oxide Films

Like most nanomanufacturing processes, molecular beam epitaxy (MBE) processes are based on atomic-level control of growing films and thus are sensitive to subtle changes that make repeatability and reproducibility of desired performance indicators a nontrivial task. The gamut of challenges include insufficient understanding of atomic-level interactions, involvement of a large number of candidate process variables, lack of direct observation and measurement techniques for key performance indicators, and significant cost and time requirements for conducting experiments. A conventional design of experiment-based analysis becomes an unrealistic option due to its demand on extensive experimentation. In this paper, we present a hybrid approach that combines current process knowledge, artificial neural networks, and design of experiments (DOE) to make use of preliminary experimental data to analyze the process behavior, enhance process knowledge, and lay down foundations for cost effective systematic experimentation. Based on preliminary experimental data generated while exploring the MBE process for growing a MgO interface layer on 6H-SiC substrate, we developed a neural-network-based meta model that can interpolate and estimate the process responses to any combination of process variable settings within the input space. Using the neural-network model trained on preliminary experimental data, we estimate the process responses for a three-level full-factorial DOE runs. Based on these runs, the DOE based analysis is carried out. The results help explain the MgO film growth dynamics with respect to process variables such as substrate temperature, growth time, magnesium source temperature, and trace oxygen on the initial substrate surface. This approach can be expanded to statistically analyze the dynamics of other complex nanoprocesses when only the exploratory preliminary experimental data are available. This approach can also lay the foundation for efficient and systematic experimentation to further analyze and optimize the processes to address issues such as process repeatability and reliability.

CAPSULE: AN INNOVATIVE CAPSTONE-BASED PEDAGOGICAL APPROACH TO ENGAGE HIGH SCHOOL STUDENTS IN STEM LEARNING

School children in general and high school students, in particular more often than not lose interest in STEM (science, technology, engineering, and math) education. Underrepresented and female students are even more discouraged by STEM courses. Our investigation and interviews with high school teachers cite that the main reason for such disinterest is the disconnect between school and reality. Students cannot relate the abstract concepts they learn in physics, biology, chemistry, or math to their surroundings. This paper discusses a new capstone project-based approach that closes this gap. This work is an outcome of an NSF funded project called CAPSULE (Capstone Unique Learning Experience). We use the top-down pedagogical approach instead of the traditional bottom-up approach. The top-down approach relates the abstract concepts to exciting open-ended capstone projects where students are engaged in designing solutions, like products to solve open-ended problems. This top-down approach is modeled after the college-level capstone design courses. The paper presents the model, its details, and implementation. It also presents the formative and summative evaluation of the model after deploying it in the Boston Public Schools, a system heavily populated by the targeted student groups.Copyright

Assessement of current health of hard disk drives

After investigating several of different degradation signatures that can potentially characterize aging and failure of computer hard disk drives (HDDs), we identified that reported uncorrect, hardware ECC recovered and read write rate parameters can provide good degradation signature for assessing the condition and remaining useful life of HDDs. Using these signatures as inputs, we develop a neural network model to assess the current health of a HDD. We collected extensive data by conducting experiments on 13 HDDs in an accelerated degradation mode. Experiments on 13 HDDs generated several hundreds of data points during their operating life. We used two thirds of these data points for computing the neural network parameters and the rest for evaluating the accuracy of model predictions. The results indicate that the trained neural network is able to assess the health of a HDD correctly 88 times out of 100 instances.

Phase Synchronization Approach to Construction and Analysis of Stock Correlation Network

Abstract Stock correlation network, a subset of financial network, is built on the stock price correlation. It is used for observing, analyzing and predicting the stock market dynamics. Existing correlation methods include the minimum spanning tree (MST), planar maximally filtered graph (PMFG), and winner take all (WTA). The MST and PMFG methods lose information due to the connection criterion and thereby fail to include certain highly correlated stocks. The WTA method, when used for a non-linear system such as stock prices, fails to capture the dynamic behavior embedded in the time series of the stocks. In this paper we present a new method, which we call phase synchronization (PS) for constructing and analyzing the stock correlation network. The PS method captures the dynamic behavior of the time series of stocks and mitigates the information loss. To test the proposed PS method we use the weekly closing stock prices of the S&P index (439 stocks) from 2000-2009. The PS method provides valuable insights into the behavior of highly correlated stocks which can be useful for making trading decisions. The network exhibits a scale free degree distribution for both chaotic and non-chaotic periods.

A Q-learning-based adaptive grouping policy for condition-based maintenance of a flow line manufacturing system

There has been a considerable progress in condition-based maintenance (CBM) in which maintenance actions are carried out as warranted by the condition of machines to reduce the associated maintenance costs and increase the availability of machines. If the maintenance activities are carried out individually, setup costs would be higher and the system downtime would be longer than if the maintenance activities are carried out together on a group of machines. So, finding an optimal grouping policy is an important problem in itself. This paper investigates a Q-learning algorithm to come up with a grouping policy that would reduce set up costs and increase the uptime efficiency of a flow line manufacturing system. The breakdown of even a single machine in a flow line system could affect the availability of the entire system, particularly when there are no storage buffers in between successive machines. The results reported here show that proposed Q-learning-based grouping policy is capable of reducing the number of repair or maintenance interruptions considerably.

Assessment of Current Health of Hard Disk Drives

Prognostics and health management of computer hard disk drives is beneficial from two different angles: it can help computer users plan for timely replacement of HDDs before they catastrophically fail and cause serious data loss; it can also help product recover facilities reuse hard disks recovered from the end-of-life computers for building refurbished computers. This paper presents a HDD health assessment method using Self-Monitoring, Analysis, and Reporting Technology (SMART) attributes. It also presents the state-of-the art results in monitoring the condition of hard disks and offers future directions for distributed hard disk monitoring.Copyright