Teresa Wu

A model for inbound supply risk analysis

Managing risk has become a critical component of supply chain management. The implications of supply chain failures can be costly and lead to significant customer delivery delays. Though, different types of supply chain vulnerability management methodologies have been proposed for managing supply risk, most offer only point-based solutions that deal with a limited set of risks. This research aims to reinforce inbound supply chain risk management by proposing an integrated methodology to classify, manage and assess inbound supply risks. The contributions of this paper are four-fold: (1) inbound supply risk factors are identified through both an extensive academic literature review on supply risk literature review as well as a series of industry interviews; (2) from these factors, a hierarchical risk factor classification structure is created; (3) an analytical hierarchy processing (AHP) method with enhanced consistency to rank risk factor for suppliers is created; and (4) a prototype computer implementation system is developed and tested on an industry example.

Methodology for supply chain disruption analysis

Given the size, complexity and dynamic nature of many supply chains, there is a need to understand the impact of disruptions on the operation of the system. This paper presents a network-based modelling methodology to determine how changes or disruptions propagate in supply chains and how those changes or disruptions affect the supply chain system. Understanding the propagation of disruptions and gaining insight into the operational performance of a supply chain system under the duress of an unexpected change can lead to a better understanding of supply chain disruptions and how to lessen their effects. The modelling approach presented, Disruption Analysis Network (DA_NET), models how changes disseminate through a supply chain system and calculates the impact of the attributes by determining the states that are reachable from a given initial marking in a supply chain network. This ability will permit better management of the supply chain and thus will allow an organization to offer quicker response times to the customer, lower costs throughout the chain, and to the end customer higher levels of flexibility and agility, lower inventories throughout the chain (both with work-in-process and inventories), lower levels of obsolescence and a reduced bullwhip effect throughout the chain. This is of particular benefit in large-scale systems, since it can give the user the ability to perform detailed analysis of a dynamic system without the computational burden of a full-scale execution of the model. Consequently, the model may then be segmented to evaluate only the portions or sub-networks that are affected by changes in an initial marking.

Learning brain connectivity of Alzheimer's disease by sparse inverse covariance estimation

Rapid advances in neuroimaging techniques provide great potentials for study of Alzheimers disease (AD). Existing findings have shown that AD is closely related to alteration in the functional brain network, i.e., the functional connectivity between different brain regions. In this paper, we propose a method based on sparse inverse covariance estimation (SICE) to identify functional brain connectivity networks from PET data. Our method is able to identify both the connectivity network structure and strength for a large number of brain regions with small sample sizes. We apply the proposed method to the PET data of AD, mild cognitive impairment (MCI), and normal control (NC) subjects. Compared with NC, AD shows decrease in the amount of inter-region functional connectivity within the temporal lobe especially between the area around hippocampus and other regions and increase in the amount of connectivity within the frontal lobe as well as between the parietal and occipital lobes. Also, AD shows weaker between-lobe connectivity than within-lobe connectivity and weaker between-hemisphere connectivity, compared with NC. In addition to being a method for knowledge discovery about AD, the proposed SICE method can also be used for classifying new subjects, which makes it a suitable approach for novel connectivity-based AD biomarker identification. Our experiments show that the best sensitivity and specificity our method can achieve in AD vs. NC classification are 88% and 88%, respectively.

MMR: An algorithm for clustering categorical data using Rough Set Theory

A variety of cluster analysis techniques exist to group objects having similar characteristics. However, the implementation of many of these techniques is challenging due to the fact that much of the data contained in todays databases is categorical in nature. While there have been recent advances in algorithms for clustering categorical data, some are unable to handle uncertainty in the clustering process while others have stability issues. This research proposes a new algorithm for clustering categorical data, termed Min-Min-Roughness (MMR), based on Rough Set Theory (RST), which has the ability to handle the uncertainty in the clustering process.

An Adaptive Particle Swarm Optimization With Multiple Adaptive Methods

Particle swarm optimization (PSO) has attracted much attention and has been applied to many scientific and engineering applications in the last decade. Most recently, an intelligent augmented particle swarm optimization with multiple adaptive methods (PSO-MAM) was proposed and was demonstrated to be effective for diverse functions. However, inherited from PSO, the performance of PSO-MAM heavily depends on the settings of three parameters: the two learning factors and the inertia weight. In this paper, we propose a parameter control mechanism to adaptively change the parameters and thus improve the robustness of PSO-MAM. A new method, adaptive PSO-MAM (APSO-MAM) is developed that is expected to be more robust than PSO-MAM. We comprehensively evaluate the performance of APSO-MAM by comparing it with PSO-MAM and several state-of-the-art PSO algorithms and evolutionary algorithms. The proposed parameter control method is also compared with several existing parameter control methods. The experimental results demonstrate that APSO-MAM outperforms the compared PSO algorithms and evolutionary algorithms, and is more robust than PSO-MAM.

Heterogeneous data fusion for alzheimer's disease study

Effective diagnosis of Alzheimers disease (AD) is of primary importance in biomedical research. Recent studies have demonstrated that neuroimaging parameters are sensitive and consistent measures of AD. In addition, genetic and demographic information have also been successfully used for detecting the onset and progression of AD. The research so far has mainly focused on studying one type of data source only. It is expected that the integration of heterogeneous data (neuroimages, demographic, and genetic measures) will improve the prediction accuracy and enhance knowledge discovery from the data, such as the detection of biomarkers. In this paper, we propose to integrate heterogeneous data for AD prediction based on a kernel method. We further extend the kernel framework for selecting features (biomarkers) from heterogeneous data sources. The proposed method is applied to a collection of MRI data from 59 normal healthy controls and 59 AD patients. The MRI data are pre-processed using tensor factorization. In this study, we treat the complementary voxel-based data and region of interest (ROI) data from MRI as two data sources, and attempt to integrate the complementary information by the proposed method. Experimental results show that the integration of multiple data sources leads to a considerable improvement in the prediction accuracy. Results also show that the proposed algorithm identifies biomarkers that play more significant roles than others in AD diagnosis.

Supplier evaluation and selection: an augmented DEA approach

Evaluating and selecting suppliers is an essential part of effectively managing todays dynamic and global supply chains. In this paper, we propose a supplier evaluation and selection methodology based on an extension of data envelopment analysis (DEA) that can evaluate suppliers in an efficient manner. Through the incorporations of a range of virtual standards, the proposed methodology termed augmented DEA, has enhanced discriminatory power over basic DEA models to rank suppliers. In addition, weight constraints are introduced to reduce the possibility of having inappropriate input and output factor weights. We demonstrate the application of augmented DEA with comparison experiments and find that the augmented DEA model has advantages over the basic DEA model as well as the cross-efficiency and super-efficiency models. Finally, we present a case application with data obtained from a communication and aviation electronics company to demonstrate the applicability and use of augmented DEA.

AIDEA: a methodology for supplier evaluation and selection in a supplier-based manufacturing environment

Supplier evaluation and selection, an important element in supplier-based manufacturing and supply chain management has been gaining attention in both academic literature and industrial practice. In this paper, we presented a modified data envelopment analysis (DEA) method for supplier selection which can operate under conditions of imprecise information. A brief description of the importance of supplier evaluation and selection is given, followed by an overview of DEA, along with several extensions of DEA applied specifically to supplier evaluation and selection. Contributions of our proposed approach include the elimination of the poor discriminatory power and inability of traditional DEA to rank the efficient suppliers. In addition, our method can be used in situations with imprecise data.

Network-based approach to modelling uncertainty in a supply chain

Supply chains are interlinked networks of suppliers, manufacturers, distributors and customers that provide a product or service to customers. Typical supply chains can be characterized by their complexity and by the inherent uncertainty in their operations. Therefore, modelling such supply chains is a difficult and challenging research task, particularly given the need to model the stochastic operations of typical supply chains. What is giving added urgency to the need to address this issue are the recent developments in communications, primarily based on Internet technologies, that offer the promise of connecting suppliers, assemblers and customers in a seamless network of information. This offers the promise of substantially improved decision-making and a consequent considerable improvement in operations. However, fulfilment of this promise is dependent on the development of a suitable modelling methodology for supply chains. A network-based methodology to model and analyse supply chain systems is proposed. The methodology represents the operation of a supply chain as an abstracted network. The approach allows for the inclusion of stochastic variables so that uncertainty in the operation of a supply chain can be modelled. The use of the methodology is illustrated using a case study based on company data. The contribution of this paper is threefold. First, an approach is presented that can represent the complex operation of a supply chain as an abstracted network. Second, the use of stochastic variables in this approach is described. The stochastic variables represent the uncertainty present in typical supply chains. Third, a case study is presented that illustrates how this approach can be used to improve the operation of a supply chain.

Measuring glomerular number and size in perfused kidneys using MRI

The goal of this work was to nondestructively measure glomerular (and thereby nephron) number in the whole kidney. Variations in the number and size of glomeruli have been linked to many renal and systemic diseases. Here, we develop a robust magnetic resonance imaging (MRI) technique based on injection of cationic ferritin (CF) to produce an accurate measurement of number and size of individual glomeruli. High-field (19 Tesla) gradient-echo MR images of perfused rat kidneys after in vivo intravenous injection of CF showed specific labeling of individual glomeruli with CF throughout the kidney. We developed a three-dimensional image-processing algorithm to count every labeled glomerulus. MRI-based counts yielded 33,786 ± 3,753 labeled glomeruli (n = 5 kidneys). Acid maceration counting of contralateral kidneys yielded an estimate of 30,585 ± 2,053 glomeruli (n = 6 kidneys). Disector/fractionator stereology counting yielded an estimate of 34,963 glomeruli (n = 2). MRI-based measurement of apparent glomerular volume of labeled glomeruli was 4.89 × 10(-4) mm(3) (n = 5) compared with the average stereological measurement of 4.99 × 10(-4) mm(3) (n = 2). The MRI-based technique also yielded the intrarenal distribution of apparent glomerular volume, a measurement previously unobtainable in histology. This work makes it possible to nondestructively measure whole-kidney glomerular number and apparent glomerular volumes to study susceptibility to renal diseases and opens the door to similar in vivo measurements in animals and humans.