Browsing by Author "Lin, SY"

Now showing 1 - 20 of 33

A computationally efficient method for nonlinear multicommodity network flow problems
(1997-07-01)
Lin, SY
;
Lin, CH
;
交大名義發表
;
電控工程研究所
;
National Chiao Tung University
In this paper, we present a new method for solving nonlinear multicommodity network flow problems with convex objective functions. This method combines a well-known projected Jacobi method and a new dual projected pseudo-quasi-Newton (DPPQN) method which solves multicommodity flow quadratic subproblems induced in the projected Jacobi method. The DPPQN method is a dual Newton-type method that differs very much from the conventional Lagrangian Newton method; our method fully exploits the structural advantages of network-type linear equality constraints to obtain a constant sparse approximate Hessian matrix with a decoupling structure and includes a novel finite-iteration successive projection and (truncated) seal algorithm to resolve the difficulty caused by coupling capacity constraints. The DPPQN method also consists of two decomposition effects, the commodity decomposition effect and the are decomposition effect, which resolve the potential numerical difficulties caused by large dimensions. We show the convergence of our method including the convergence of the finite-iteration successive projection and (truncated) seal algorithm. Compared with the Frank-Wolfe with PARTAN algorithm in which a price-directive decomposition method is used to solve linearized multicommodity flow problems, our method is dramatically faster in terms of the CPU time on a Sparc-10 workstation at solving numerous nonlinear multicommodity network flow examples. (C) 1997 John Wiley & Sons, Inc.
A computationally efficient method for nonlinear multicommodity network flow problems
(1997-07-01)
Lin, SY
;
Lin, CH
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, we present a new method for solving nonlinear multicommodity network flow problems with convex objective functions. This method combines a well-known projected Jacobi method and a new dual projected pseudo-quasi-Newton (DPPQN) method which solves multicommodity flow quadratic subproblems induced in the projected Jacobi method. The DPPQN method is a dual Newton-type method that differs very much from the conventional Lagrangian Newton method; our method fully exploits the structural advantages of network-type linear equality constraints to obtain a constant sparse approximate Hessian matrix with a decoupling structure and includes a novel finite-iteration successive projection and (truncated) seal algorithm to resolve the difficulty caused by coupling capacity constraints. The DPPQN method also consists of two decomposition effects, the commodity decomposition effect and the are decomposition effect, which resolve the potential numerical difficulties caused by large dimensions. We show the convergence of our method including the convergence of the finite-iteration successive projection and (truncated) seal algorithm. Compared with the Frank-Wolfe with PARTAN algorithm in which a price-directive decomposition method is used to solve linearized multicommodity flow problems, our method is dramatically faster in terms of the CPU time on a Sparc-10 workstation at solving numerous nonlinear multicommodity network flow examples. (C) 1997 John Wiley & Sons, Inc.
A hybrid classification tree for products of complicated machines in flexible manufacturing systems
(2005)
Horng, SC
;
Lin, SY
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, we propose a Hybrid Classification Tree (HCT) to classify the products of complicated machines in flexible manufacturing systems. The HCT combines a proposed clustering algorithm with the Classification and Regression Tree (CART) to take the advantage of the constant property of control settings during any process step for a type of product. The proposed clustering algorithm split the data set into terminal clusters using splitting attributes based on a separation matrix and fuzzy rules. The terminal clusters which consist of the data of more than one product will be further classfied using the CART We have tested the HCT on the products of an ion implanter for the vast number of wafers of 26 recipes and compared the classification results and the computation time with the existing software See5 and CART The comparison results show that the HCT retains the classification accuracy of CART while saving 40% training time.
A more general dual projected pseudo quasi-Newton method for large network optimization problems
(1999)
Lin, SY
;
Lin, SS
;
電控工程研究所
;
Institute of Electrical and Control Engineering
One of the major computational steps in the dual projectd pseudo quasi-Newton method for neweork type optimization problems is the projection onto a convex set formed by the linear inequlity constraints of the problem. When the inequlity constraints are coupled, the large-dimension projection problem is a difficult one. In this paper, we propose a graphic-method based decomposition technique to determine a smallest subset of inequality contraints to be converted to equality constraints so that the rest of the inequality constraints become disjoint subsets. Consequently, the large-dimension projection problem can be decomposed into several independent small-dimension projection subproblems, and the computation complexity for the projection step will be largely reduced. Copyright (C) 1998 IFAC.
A new dual-type method used in solving optimal power flow problems
(1997-11-01)
Lin, CH
;
Lin, SY
;
交大名義發表
;
電控工程研究所
;
National Chiao Tung University
In the framework of SQP method for OPF problems, we propose a new dual-type method for solving the QP subproblems induced in the SQP method. Our method achieves some attractive features; it is computationally efficient and numerically stable. The computational formulae of our method are simple, concise and easy to be programmed. We have tested our method for OPF problems on several power systems including a 2500-bus system.
A new method for constructing fuzzy decision trees and generating fuzzy classification rules from training examples
(2000-10-01)
Chen, SM
;
Lin, SY
;
資訊工程學系
;
Department of Computer Science
This paper presents a new method for constructing fuzzy decision trees and generating fuzzy classification rules from training instances using compound analysis techniques. The proposed method call generate simpler fuzzy classification rules and has a better classification accuracy rate than the existing method. Furthermore. the proposed method generated less fuzzy classification rules.
A novel current-injected sensorless control scheme for squirrel-cage induction motors
(2002)
Lin, SY
;
Tzou, YY
;
電控工程研究所
;
Institute of Electrical and Control Engineering
This paper presents a novel sensorless control scheme of squirrel-cage induction motors by using a PWM waveform injection technique for the estimation of rotor dependent stator leakage inductances. To increase the signal-to-noise ratio when operating in low-speed region, the magnitude of the injected signal is adaptively adjusted according to motor operating condition. Practical realization issues of the detection of the reflected signals and estimation of the rotor position have been addressed. The developed sensorless control scheme has been implemented by using a high-performance single-chip DSP controller. Experimental verification of the developed adaptive injection scheme and rotor position estimation scheme has been carried out. Experimental results show the proposed sensorless position control scheme can reach a position control resolution of 0.05 rad with a rated holding torque.
A parallel block scaled gradient method with decentralized step-size for block additive unconstrained optimization problems of large distributed systems
(2003-03-01)
Lin, SY
;
Lin, SS
;
電控工程研究所
;
Institute of Electrical and Control Engineering
In this paper, we propose a modified parallel block scaled gradient method for solving block additive unconstrained optimization problems of large distributed systems. Our method makes two major modifications to the typical parallel block scaled gradient method: First, we include a pre-processing step which reduces the computational time; second, we propose a decentralized Armijo-type step-size rule. This rule circumvents the difficulty of determining a step-size in a distributed computing environment and enables the proposed parallel algorithm to execute in a distributed computer network with a limited amount of data transfer. We have applied our method to the weighted-least-square problems of power system state estimation and demonstrated the convergence of our method by testing numerous examples on a PC network. The speedup ratio of the distributed version of our method tends to increase proportionally with the number of subsystems (or computers).
A power-spectrum based neural fuzzy connection admission mechanism for ATM networks
(1997)
Chang, CJ
;
Lin, SY
;
Shiue, YR
;
Cheng, RG
;
交大名義發表
ATM (asynchronous transfer mode) networks support services with bursty traffic. A sophisticated and real-time admission controller is needed not only to guarantee the required quality-of-service (QoS) for existing calls but also to achieve higher system efficiency. In this paper, we propose a power-spectrum based neural fuzzy connection admission control (PNFCAC) mechanism for an ATM network. Power-spectrum of input process explicitly describes the correlation behavior of input traffic and has a great impact on the queueing behavior of an ATM net work. The neural fuzzy controller has been successfully applied to deal with the complex problem which can not be analyzed by classical methods based on probability theory and bivalent logic for recent years. Under the constraint of QoS, we construct a decision hyperplane of the connection admission control according to parameters of the power spectrum. We further adopt the learning/adapting capabilities of the neural network to adjust the optimum location of the boundary between these two decision space. Simulation results show that the PNFCAC mechanism provides a superior system utilization over the conventional CAC schemes and the neural fuzzy CAC using time-domain analysis, while keeping the QoS contract.
ACTAM: Cooperative multi-agent system architecture for urban traffic signal control
(2005-01-01)
Chen, RS
;
Chen, DK
;
Lin, SY
;
資訊管理與財務金融系註：原資管所+財金所
;
Department of Information Management and Finance
The traffic congestion problem in urban areas is worsening since traditional traffic signal control systems cannot provide] efficient traffic regulation. Therefore, dynamic traffic signal control in Intelligent Transportation System (ITS) recently has received increasing attention. This study devised a multi-agent architecture, the Adaptive and Cooperative Traffic light Agent Model (ACTAM), for a decentralized traffic signal control system. The proposed architecture comprises a data Storage and communication layer, a traffic regulation factor processing layer, and a decision-making layer. This study focused on utilizing the cooperation of multi-agents and the prediction mechanism of our architecture, the Forecast Module, to forecast future traffic volume in each individual intersection. The Forecast Module is designed to forecast traffic volume in an intersection via multi-agent cooperation by exchanging traffic volume information for adjacent intersections, since vehicles passing through nearby intersections were believed to significantly influence the traffic volume of specific intersections. The proposed architecture can achieve dynamic traffic signal control. Thus, total delay time of the traffic network under ACTAM can be reduced by 37% compared to the conventional fixed sequence traffic signal control strategy. Consequently, traffic congestion in urban areas can be alleviated by adopting ACTAM.
An effective application of decision tree to stock trading
(2006-08-01)
Wu, MC
;
Lin, SY
;
Lin, CH
;
工業工程與管理學系
;
Department of Industrial Engineering and Management
This paper presents a stock trading method by combining the filter rule and the decision tree technique. The filter rule, having been widely used by investors, is used to generate candidate trading points. These points are subsequently clustered and screened by the application of a decision tree algorithm C4.5. Compared to previous literature that applied such a combination technique, this research is distinct in incorporating the future information into the criteria for clustering the trading points. Taiwan and NASDAQ stock markets are used to justify the proposed method. Experiment results show that the proposed trading method outperforms both the filter rule and the previous method. (C) 2005 Elsevier Ltd. All rights reserved.
An effictive and efficient hierarchical fuzzy rule based classifier
(2003)
Tsai, CH
;
Lin, SY
;
Cheng, MH
;
Horng, SC
;
Liu, CH
In this paper, we propose a hierarchical fuzzy rule based classifier (HFRBC) for the classification problem with large number of classes and continuous attributes. A hierarchical clustering concept is introduced to achieve a finer fuzzy partition. Critical attributes are used to perform the cluster splitting and generate a cluster splitting tree. The effective attributes for the terminal clusters in the cluster splitting tree are picked so as to reduce the size of the fuzzy-rule set and hence reduce the computational complexity. The fuzzy rule generation procedures and classification procedures of the proposed HFRBC are simple and easily implemented. We have successfully applied the HFRBC to the classification problem of the working wafers in an ion implanter.
An ordinal optimization theory-based algorithm for solving the optimal power flow problem with discrete control variables
(2004-02-01)
Lin, SY
;
Ho, YC
;
Lin, CH
;
電控工程研究所
;
Institute of Electrical and Control Engineering
The optimal power flow (OPF) problem with discrete control variables is an NP-hard problem in its exact formulation. To cope with the immense computational-difficulty of this problem, we propose an ordinal optimization theory-based algorithm to solve for a good enough solution with high probability. Aiming for hard optimization problems, the ordinal optimization theory, in contrast to heuristic methods, guarantee to provide a top n% solution among all with probability more than 0.95. The approach of our ordinal optimization theory-based algorithm consists of three stages. First, select heuristically a large set of candidate solutions. Then, use a simplified model to select a subset of most promising solutions. Finally, evaluate the candidate promising-solutions of the reduced subset using the exact model. We have demonstrated the computational efficiency of our algorithm and the quality of the obtained solution by comparing with the competing methods and the conventional approach through simulations.
Architecture and operations of the VLSI array processors for implementing a receding horizon controller
(1995)
Lin, SY
;
電控工程研究所
;
Institute of Electrical and Control Engineering
Auto-tuning self-commissioning control of a universal drive
(1998)
Tzou, YY
;
Wu, H
;
Lin, SY
;
Lin, JY
;
電控工程研究所
This paper presents a systematic design procedure for the auto-tuning self-commissioning (ATSC) control of a universal drive. The ATSC controller consists of five major parts: a current vector controller (CVC), a field-oriented vector controller (FVC), a servo loop controller (SLC), a self-commissioning controller (SCC), and an auto-tuning controller (ATC). The developed universal drive can be programmed to operate as a general-purpose adjustable speed inverter drive, a vector-controlled sensorless induction drive, a de servo drive, a permanent-magnet sc servo drive, or an induction servo drive. The concept and architecture for the future development of universal drives are studied in this paper. We also illustrate some experimental studies of the developed universal drive in application to the ATSC control of an induction motor with unknown motor parameters.
Basic hardware module for a nonlinear programming algorithm and applications
(1997-08-01)
Lin, SY
;
交大名義發表
;
電控工程研究所
;
National Chiao Tung University
;
Institute of Electrical and Control Engineering
We present a VLSI-array-processor architecture for the implementation of a nonlinear programming algorithm that solves discrete-time optimal control problems for nonlinear systems with control constraints. We incorporate this hardware module with a two-phase parallel computing method and develop a VLSI-array-processor architecture to implement a receding horizon controller for constrained nonlinear systems. On the basis of current VLSI technologies, the estimated computing time to obtain the receding-horizon feedback-control solution meets the realtime processing-system needs. (C) 1997 Elsevier Science Ltd.
Distributed motion control using real-time network communication techniques
(2000)
Lin, SY
;
Ho, CY
;
Tzou, YY
;
電控工程研究所
;
Institute of Electrical and Control Engineering
This paper presents a survey of the applications of real-time network control technology to advanced digital motion control system. Applications of PROFIBUS, FIP, CAN bus, Interbus-S, SERCOS, and Ethernet to coordinated motion control have been studied. Realization issues in application of real-time network control for motor drives and coordinated motion control systems are discussed. Future development trend of realtime network for motion control has been forecasted.
Dual-DSP sensorless speed control of an induction motor with adaptive voltage compensation
(1996)
Tzou, YY
;
Lee, WA
;
Lin, SY
;
電控工程研究所
;
Institute of Electrical and Control Engineering
Fault detection and isolation for plasma etching using model-based approach
(2003)
Cheng, MH
;
Huan-Shin, L
;
Lin, SY
;
Liu, CH
;
Lee, WY
A fault detection and isolation system using model-based approach for the chamber pressure of plasma etching is developed. The dynamics of chamber pressure is modeled as a linear multiple-input-single-output closed-loop system and the model parameters are extracted by the system identification technique. The obtained parameters are then converted to physically meaningful features for detecting and isolating faults. The fuzzy inference and Dempster-Shafer evidence combining techniques are employed to detect and isolate fault from the features. The system has been evaluated by the measured data that are collected via SECS-II from the employed Lam-490 plasma etcher for processing practical products. The test results are satisfactory for a large amount of practical data and extensive computer simulations.
Fuzzy-tuning current-vector control of a three-phase PWM inverter for high-performance AC drives
(1998-10-01)
Tzou, YY
;
Lin, SY
;
電控工程研究所
;
Institute of Electrical and Control Engineering
This paper proposes a new discrete fuzzy-tuning current-vector control (FTC) scheme for three-phase pulsewidth modulation (PWM) inverters. The proposed current control scheme can achieve fast transient responses and, at the same time, have very low total harmonic distortion in output current during steady-state operation. The proposed FTC scheme generates quasi-optimum PWM patterns by using a closed-loop control technique with instantaneous current feedback. The proposed PTC scheme has been realized using a single-chip digital signal processor (TMS320C14) from Texas Instruments. Experimental results are given to verify the proposed fuzzy-tuning current control strategy for three-phase PWM inverters.