Xindu Chen

Optimal configuration of cluster supply chains with augmented Lagrange coordination

A distributed CSCC model enabling coupled decision mechanism is built with ALC formulation.A systematic ALC approach comprising analysis, model, and solution techniques is proposed.A generic ALC solution framework and procedure is proposed toward CSCC problems.An experiment and sensitivity analysis are provided to show effectiveness of using ALC for resolving a CSCC problem. Enterprises in an industrial cluster could dynamically alliance in the form of cluster supply chains to share inner-cluster resources and services, and respond to the ever-fluctuating customer demands in a cost-effective way. However, an effective and feasible method enabling such dynamic cluster supply chain configuration (CSCC) lags behind practice due to the conflict of interests. Researchers are designing All-in-One theoretic models to optimize CSCC with the assumed decision details of all enterprises, while in fact clustered enterprises are seeking effective decentralized decision mechanisms which protect their decision autonomy in the frequently re-configured CSC. A newly emerged multi-disciplinary optimization method, Augmented Lagrangian Coordination (ALC), which supports the open-structure collaboration with strict optimization convergence, is thoroughly investigated in this paper and applied to solve the conflict. Through a complete analysis of CSCs configuration policies in typical stages, a generic CSCC model is proposed and then partitioned into an ALC-based decentralized decision model by the typical decision autonomy distribution in clusters. Clustered enterprises collaborate vertically and laterally along the ALC model through multi-dimensional couplings to achieve the overall consistency and optimality. Results have proved the effectiveness of ALC for CSCC problem. A set of sensitivity analysis is also conducted to find out the condition in which an order has to be fulfilled in a CSC and the most appropriate configuration.

Extending analytical target cascading for optimal supply chain network configuration of a product family

Platforming is not only a powerful approach to new product development by sharing common and modular components and processes but also allows the supply chain to gain benefits from risk-pooling effect through shared resources. Analytical target cascading (ATC) is a decentralised method suitable for configuring a hierarchical supply chain of an assembled product while accommodating necessary degree of decision autonomy and information privacy of individual enterprises. Because product variants in a family share platform components, the supply chain structure becomes a weakly networked hierarchy where a small number of elements are laterally linked. In addition, multiple customers who share a common platform component may require the corresponding supplier to use different strategies such as just-in-time (JIT) and lowest price to supply the component. As a result, different decision variables may be involved in the interaction between a shared lower-level element and its different parental elements. This paper develops a new ATC method suitable for dealing with these two special characteristics in supply chain configuration (SCC) for a product family. Numerical results demonstrate that the new method produces better results than those obtained from using the ATC method that only allows a supplier to optimise the supply of the platform component to all its customers with the same strategy.

Qualitative search algorithms for partner selection and task allocation in the formulation of virtual enterprise

In the formation of dynamic virtual enterprise (VE), partner search is of prime importance. Finding potential partners in a directory is the first step of the process that might involve several interactions between the VE initiator and these potential partners in order to gather additional information and to reach agreements on the forms and terms of cooperation (Camarinha-Matos and Afsarmanesh, 1999, pp. 493). Given the manufacturing requirements, the initiator or coordinator of dynamic VE should first prepare several alternative schemes that can satisfy the requirements. Next, the coordinator should evaluate these alternative schemes by means of some multi-objective analysis methodologies, such as multi-objective decision-making (MODM), data envelopment analysis (EDA), analytic hierarchy process (AHP) or fuzzy analytic hierarchy process (FAHP), fuzzy comprehensive evaluation (FCE), or statistical analysis, multi-attribute utility theory (MAUT). Then the coordinator should determine the formation scheme. The present paper deals with the search procedure of alternative schemes to form VE for the coordinator. In this paper, a logic model is first put forward to describe the relationships among the manufacturing capabilities of enterprises and the manufacturing requirements of clients within the range of the applied industry. It will suffer little change in a considerable long period because of its qualitative nature. This advantage will make it easy to maintain. Next, the procedures for the manufacturing enterprises seeking partners with this logic qualitative model are expressed mathematically. Based upon this logic model, three search algorithms are brought about for three different optimal goals, respectively: minimizing the number of enterprises, minimizing the number of redundant basic capability units in the dynamic enterprise alliances and minimizing the number of basic capability units useful to the manufacturing requirement of each enterprise. After that, theoretical analyses are made for these algorithms. It is demonstrated that these three search algorithms are all polynomial with respect to the number of enterprises belonging to a particular industry. This number is usually fairly large and represents the size of the search problem. In the end, it is shown that these three algorithms are effective and reliable with an example.

RFID-enabled smart assembly workshop management system

Although RFID technologies are enjoying rapid developments, its practical use in manufacturing processes is still limited. Existed applications focus mainly on the pallet-level RFID implementation, which facilitates batch data processing in material handling operations, e.g. material delivery, yet cannot realize real-time lean control for individual parts. This leaves the manufacturing processes as a blank area for taking full advantages of RFID. This paper discusses the item-level RFID implementation in terms of both real-time information control mechanism and system development. A simulated automobile assembly line is employed for concept demonstration. Three questions will be addressed. First, how to use RFID systems to enable the real-time coordination and interaction between the production planning and execution levels to achieve the lean control of manufacturing processes. Second, how to realize the RFID-enabled smart management for typical manufacturing processes, including assembling, packaging, buffering, etc. Thirdly, how to establish a real-time information infrastructure to integrated the typical RFID-enabled smart processes.

Image super-resolution via a novel cascaded convolutional neural network framework

Abstract Due to excellent self-learning ability, deep convolutional neural networks (CNNs) are successfully employed in the field of single image super-resolution (SR) compared with interpolation methods and sparse coding methods. To implement the multi-scale image SR task with a single trained model and to further improve the performance of image SR, we propose a novel cascaded CNN framework with three stages, which are feature extraction, detail prediction and reconstruction. At the stage of feature extraction, we propose a scheme of multi-scale feature mapping to extract the inherent features via the low-resolution image. At the stage of detail prediction, the lost details of the original low-resolution image are predicted via the network cascade. Finally, a high-resolution image is achieved by the scheme of residual learning, which is implemented by superimposing the lost details and the original low-resolution image. To avoid gradient explosions, we use gradient clipping to train the proposed cascaded CNN framework. Comparison results indicate that our proposed cascaded CNN framework for image SR is superior to many state-of-the-art methods.

An Effective Optimization Scheme for ECG Signal Denoising via Low-Rank Matrix Decomposition

Electrocardiogram (ECG) signal denoising is an important preprocessing for ECG signal analysis. The contaminated ECG signal can be considered as the combination of the desired clean signal and the noise. Thus, ECG signal denoising can be considered as a problem of obtaining an optimal solution to the desired clean signal. In this paper, an effective optimization scheme for ECG signal denoising is presented based on low-rank matrix decomposition. First, the ECG denoising problem is formulated as low-rank matrix decomposition. So, an ECG beats matrix is assumed to be the combination of a sparse noise matrix and a low-rank matrix. Considering the repeatability of ECG signal, the rank of the ECG beats matrix is assumed to be one. Then, to fully exploit the low-rank property of the ECG signal, the matrix decomposition is modified by means of adding different weights to different singular values. Finally, the desired clean ECG signal is reconstructed by the low-rank component. The experimental results show that the proposed denoising method achieves the best performance of suppressing the electromyographic noise in the ECG signals compared with other optimization models.

Fabrication and evaluation of controllable deposition distance for aligned pattern by multi-nozzle near-field electrospinning

To mass-volume fabricate micro- and nano-scales aligned pattern, multi-nozzle near-field electrospinning (NFES) direct-writing technology is well proposed as a high-efficiency method in electrohydrodynamic (EHD) printing process. However, the interference effect among adjacent nozzles and coupling effect of various parameters have restricted to investigate deposition characteristic of multi-nozzle NFES and control EHD multi-jet deposition accuracy. In order to improve the accuracy of EHD multi-jet deposition with high-efficiency printing process, the experimental result compared with theoretical method were discussed. In this work, the influence of multi-nozzle geometry distribution and electrospinning parameters on deposition characteristic was studied with multi-nozzle NFES setup, and nozzles were in linear array. The deposition distance and homogeneity of aligned nanofibers were measured and explained with coefficient of dispersion on electric field among nozzles by simulation. Moreover, deposition distance of multi-nozzle NFES process was evaluated by modified theoretical derivation based on our previous studies. The modified theoretical derivation showed a good agreement with experiment results, and indicated that multi-nozzle NFES could accurately and efficiently direct-write aligned array pattern in future.To mass-volume fabricate micro- and nano-scales aligned pattern, multi-nozzle near-field electrospinning (NFES) direct-writing technology is well proposed as a high-efficiency method in electrohydrodynamic (EHD) printing process. However, the interference effect among adjacent nozzles and coupling effect of various parameters have restricted to investigate deposition characteristic of multi-nozzle NFES and control EHD multi-jet deposition accuracy. In order to improve the accuracy of EHD multi-jet deposition with high-efficiency printing process, the experimental result compared with theoretical method were discussed. In this work, the influence of multi-nozzle geometry distribution and electrospinning parameters on deposition characteristic was studied with multi-nozzle NFES setup, and nozzles were in linear array. The deposition distance and homogeneity of aligned nanofibers were measured and explained with coefficient of dispersion on electric field among nozzles by simulation. Moreover, deposition dis...

Modelling and prediction of the effect of cutting strategy on surface generation in ultra-precision raster milling

This paper studies the effect of cutting strategy on surface generation in ultra-precision raster milling (UPRM). By adding the influences of shift length and tool-interference on surface generation, a holistic surface roughness prediction model is built which takes into account the effect of cutting parameters, tool path generation, geometry parameters of diamond tool, the size of the workpiece and machine characteristics. The optimal shift ratio can be achieved by changing factors involved in developing cutting strategy to improve surface quality without decreasing machining efficiency. Conditions for the presence of tool-interference in UPRM are presented. Based on the holistic surface generation model, an integrated system is developed to automatically generate the numerical control program, and predict surface quality and machining efficiency. A series of cutting experiments has been conducted to verify the proposed surface generation model and test the performance of the integrated system. The experimental results agree well with the predicted results from the model and the integrated system.