Off-chip buses consume a huge fraction (20%-40%) of the system power. Hence, techniques such as increasing bus widths, transition encoding etc. have been used for power reduction on off-chip data buses. Since capacitances at the I/O pads and interwire capacitances contribute significantly to increase in power, encoding/decoding schemes have been developed to reduce switching activity of the off-chip bus lines, thus reducing power. Frequent-Value Encoding(FVE) [1], Frequent Value Encoding with Xor (FVExor) [1] and VALVE [2] are some of the better known encoding schemes but they still have scope for improvement. This thesis addresses the problem of power reduction in off-chip data buses by encoding variable number (1 to 4) of fixed-size (32-bit) data values (variable length patterns) which exhibit temporal locality. This characteristic enables us to cache these patterns using 64-entry CAM at the encoder and 64-entry SRAM at the decoder. Whenever a pattern match occurs a 2-bit code indicating the index of the match is sent. If a variable length pattern match occurs then the code and unmatched portion of data is sent. We implemented our scheme, Variable Length Pattern Coding (VLPC) for various integer and floating point benchmarks and have seen 6% to 49% encodable patterns in these benchmarks. Based on the experiments on simplescalar and our analysis in MATLAB, we obtained 4.88% to 40.11% reduction in transition activity for SPEC2000 benchmarks such as crafty, swim, mcf, applu, ammp etc. over unencoded data. This is 0.3% to 38.9% higher than that obtained using FVE, FVExor [1] and VALVE [2] encoding schemes. Finally, we have designed a low-power custom CAM and SRAM using 45nm BSIM4 technology models which has been used to verify lower latency of data matching and storing.

Variable frequency drives typically have employed dc voltage or current links for power distribution between the input and output converters and as a means to temporarily store energy. The dc link based power conversion systems have several inherent limitations. One of the important limitations is the high switching loss and high device stress which occur during switching intervals. This severely reduces the practical switching frequencies. Additionally, while the cost, size, and weight of the basic voltage sourced PWM drive is attractive, difficulties with input harmonics, output dV/dt and over-voltage, EMI/RFI, tripping with voltage sags, and other problems significantly diminish the economic competiveness of these drives. Add-ons are available to mitigate these problems, but may result in doubling or tripling the total costs and losses, with accompanying large increases in volume and weight. This research investigates the design, control, operation and efficiency calculation of a new power converter topology for medium and high power ac-ac, ac-dc and dc-ac applications. An ac-link formed by an inductor-capacitor pair replaces the conventional dc-link. Each leg of the converter is formed by two bidirectional switches. Power transfer from input to output is accomplished via a link inductor which is first charged from the input phases, then discharged to the output phases with a precisely controllable current PWM technique. Capacitance in parallel with the link inductor produces low turn-off losses. Turn-on is always at zero voltage as each switch swings from reverse to forward bias. Reverse recovery is with low dI/dt and also is buffered due to the link capacitance.

The manufacturing process has been a heavily studied area over the past century. The study completed herein has established a foundation for the future of manufacturing research. The next step of this industry is to become proficient at the micro and nano scale levels of manufacturing. In order to accomplish this goal, the modeling of machining system needs to be completely understood throughout the entire process. In effort to attack this problem, this study will focus on the boundaries present in machining systems; and will define and interpret the associated phenomena. This particular focus is selected since nearly all manufacturing related studies concentrate on continuous processes; which by definition considers only one particular operation. There is a need to understand the phenomena corresponding to interactions of multiple processes of manufacturing systems. As a means to this end, the nonlinear phenomena associated in the continuous domains of machining systems will be modeled as linear to ensure the boundary interactions are clearly observed. Interference of additional nonlinearities is not the focus of this research. In this dissertation, the mechanical model for a widely accepted machine-tool system is presented. The state and continuous domains are defined with respect to the boundaries in this system (contact and frictional force acting at the point of tool and work-piece contact). The switching sets defining plane boundaries for the continuous systems of this machine-tool will be defined and studied herein. The forces and force products, at the point of switching from one continuous system to another, govern the pass-ability of the machine-tool through the respective boundary. The forces and force product components at the switching points are derived according to discontinuous systems theory Luo [1]. Mapping definitions and notations are developed through the switching sets for each of the boundaries. A mapping structure and notation for periodic interrupted cutting, non-cutting and chip seizure motions are defined. The interruption of the chip flow for a machining system will be investigated through a range of system parameters. The prediction of interrupted periodic cutting, non-cutting and chip seizure motion will be completed via closed form solutions for this machine-tool. The state of this system is defined to utilize the theory of Luo [1]. This is necessary to properly handle the frictional force boundary at the chip/tool interface, the onset of cutting boundary and the contact boundary between the tool and work-pieces. The predictions by this method will be verified via numerical simulation and comparison to existing research. A goal of this research is to illustrate the effects of the dynamical systems interacting at the frictional force (chip/tool) boundary and the chip onset of growth and vanishing boundary. The parameter space for this machine-tool model is studied through numerical and analytical predictions, which provide limits on the existence of interrupted periodic cutting, non-cutting and chip seizure motions.

Health pursuing is one of the most important and common aspirations of people in a modern society. However, what is their concept about health? What are the conditions of their cognitions about being healthy? As people¡¦s income has increased, nutritious diet supplements (generally called health foods) have become a popular product category. But, while types of health foods in the market are numerous and competition between the different products are fierce, how do consumers make their decision, facing these various choices? How do Taiwanese people understand and acquaint with particular types of health foods? How do they switch between them, if there is a need for switch among the particular products. The current research tried to develop a measurement scale for the concept of health suitable for Taiwanese people based on their daily life, using health confidence, health dieting, living adaptation, and social interaction as the four fundamental dimensions. This scale is different from those health concept measurement developed based on the traditional medical perspective. The major testing products of this research are ¡§Tong-Chong-Sia-Tsau¡¨ (Cordyceps Sinensis for medical terminology) and ¡§Ling-Zhi,¡¨ two of new categories of Chinese health supplements developed from special plants. Two other categories of health foods product, ¡§traditional Chinese herb medicine¡¨ and ¡§Western vitamin pills,¡¨ are also included in the observation of category switching. Adoption of innovation is an important issue in the buying behavior research of health foods. The current research developed a measurement scale of innovation adoption, adopting many measuring items from Kirton (1976). Regarding the choice behavior, this research observed and analyzed both ¡§need for cognition¡¨ and ¡§need for change¡¨ aspects, referring to Wood and Swait¡¦s (2002) choice research. The consumer samples of this research were collected from two groups of people in Taiwan: travel agents and land administration agents. The number of valid sample was 414. The relationship between health foods buying behavior and the demographic variables were as well analyzed and discussed for the valid sample.

With the rapid growth of load demand, the distribution system is becoming very complicated such that the operation efficiency and service quality are deteriorated during recent years. Engineers have to solve the problems by applying new technologies to enhance the efficiency of distribution system. In this thesis, an immune algorithm(IA) based on weighting selection as a decision maker is proposed to reach the desired switching operations such that transformer and feeder loading balance can be achieved. The IA antigen and antibody are equivalent to the objective and the feasible solution for a conventional optimization method. The concept of the information entropy is also introduced as a measure of diversity for the population to avoid falling into a local optimal solution. This algorithm prevents the possibility of stagnation in the iteration process and achieves the fast convergence for the global optimization. With the object-orient programming(OOP), this research project is to create the relationship of distribution element objects and encapsulation of data with all 22KV underground systems in Taichung district. The OOP does provide an effective tool for the management of distribution system database and the fault detection, isolation, and service restoration(FDIR) function of feeders and main transformers. According to the attributes of line switches, we can create the 22KV distribution system configuration with the topology processor. In order to calculate the current flows of line switches, this project will also execute the three phase load flow program with the customer information system(CIS), load survey, outage management information system(OMIS), and the data of all feeders and main transformers. In this thesis, the IA is used to solve the optimal switching problem by considering the customer load characteristics for the normal operation and the overload contingency of the distribution system. The efficiency of immune algorithm to solve the problem is verified by comparing to the computing time of the conventional binary integer programming for decision making of switching operation. A Taichung district distribution system is selected for computer simulation to demonstrate the effectiveness of the proposed methodology for solving the optimal switching operation of distribution system. The result of this thesis will be an important reference for distribution automation in Taiwan.

In this dissertation, the artificial intelligent Petri nets is applied to find the optimal switching operation for service restoration and feeder loading balance for 18-feeders distribution systems that containing the whole 24-hours load profiles of service zones. After the fault location has been identified and isolated for a system fault contingency, the Petri nets model with inference mechanism is derived and applied to solve the optimal load transfer among distribution feeders. For system normal operation condition, the load balancing among distribution feeders is obtained by the Petri nets model to enhance the operation efficiency of distribution systems. The switching operation, which will result in the loading balance among distribution feeders, is derived by the Petri nets model according to the loading cost of distribution systems. To determine the effectiveness of the proposed methodology, a Taipower (Taiwan Power Company) distribution system which serves a mixed types of customers is selected to perform the computer simulation. It is found that the Petri nets approach can enhance the solution process of fault restoration with proper load transfer and improve feeder load balance for distribution systems by considering the load characteristics of the service customers.

A novel non-planar and multi-reentrant two-spherical-mirror ring cavity is demonstrated. It is compact and free of astigmatism compare to the commercial ring cavity systems. The multi-reentrant condition of the ring cavity is derived and the stability of the laser cavity is analyzed. The study of polarization evolution in this kind of ring cavity is also presented. Unidirectional operation is achieved by use of reciprocal and nonreciprocal polarization rotators to differentiate the round-trip loss. The multi-reentrant ring cavity has been utilized in single frequency laser and passively Q-switched laser. Single frequency laser possesses the advantages of high coherence and low noise, which can be used to the applications such as precision measurement. In the methods of single frequency generation, ring cavity configuration was shown to be the most robust one. Using this ring cavity, an IR and its intra-cavity frequency doubled green laser were demonstrated which the amplitude noise is lower than 0.3%. Passively Q-switched laser is an efficient and compact way to generate high-peak-power laser pulses because high voltages and fast driving electronics are not required. Its high power is useful for diverse applications including nonlinear optical processes, micromachining, material processing and range finders. But the major drawback of a passively Q-switched laser is its inherent large timing jitter, which is mainly originated from the photo dynamics in the cavity, environmental instabilities and spontaneous noise from the gain medium. In our study, we demonstrated the operation of a low-jitter, passively Q-switched laser by using the reentrant two-mirror unidirectional ring cavity, which generates a pulse width of 63ns, peak power of 250 W laser output. Due to the elimination of spontaneous noise and spatial hole burning effects, the timing jitter can be maintained below 3% over a wide range of pump powers with integrations of over 52,000 pulses.

Abstract This study is to investigate the status quo and relationship between R&D-Marketing integration in the product innovation process and organizational performance. Based on the company registered in 2003, a target population composed of 259 Switching Power Supply(SPS¡^companies is formed. Questionnaires were completed by managers responsible for marketing or R&D-Marketing departments. E-mail survey were employed to collect the data and use correlation analysis, one-way ANOVA and hierarchical regression analysis to analysis the data. It is found that the organizational factors variations exist in the integration of department and have positive influence on various activities when perpetual department was established and the manager is responsible for both marketing and R&D-Marketing department. Moreover, the result of research shows organizational factors have positive influence on department integration, and the mutual participation of members is most important. Besides, organizational factors will affect financial performance and program accomplish by integration of R&D-Marketing respectively of organizational category and chief factor. Based on the research findings, the perpetual department for developing new product and co-operated managers are suggested. In addition, mutual participation of members and establishing information network and database to reduce the gap of information between two departments deserve more attention so that the new product¡¦s performance could be improved. Keywords : switching power supply , new product development, department integration, organizational performance

In a deregulated power market, customers would have more choices for their power service and the improvement of service quality has become a challenge to power transmission and distribution companies. Distribution system reliability that was traditionally considered within the planning activities, is now incorporated in the operational environment. This dissertation presents study results of a multi-objective feeder operation optimization problem that considers how to balance network efficiency, switching and reliability costs in a distribution network. The proposed method divides annual feeder load curve into multi periods of load levels and optimizes the feeder configurations for different load levels in annual operation planning. Customer load profiles and seasonal varying data of feeder section failure rates and customer interruption costs are considered. Simulations results demonstrate the time varying effects on the optimal distribution feeder reconfiguration and operation costs. A binary particle swarm optimization (BPSO) search is adopted to determine the feeder configuration in each time period. Test results indicate that not considering time varying effects and using only simplified fixed load and reliability parameters could underestimate the total loss to the utility and its customers.

In retrospect, telecommunication technology and services have seen incessant renovation and development. The wave of liberalization is also the inexorable trend in the global telecommunications industry, the telecommunications industry in Taiwan can not be excluded itself from the trend. The telecommunications industry in Taiwan has been opened by degrees and sought to establish a fair competitive environment. In the meantime, there are several important changes no matter in facets of regulatory regimes, industrial structure, technology, or market demand, etc. The environment of telecommunications industry became more volatile than the monopoly one's. We extend the opinion of Noam (2006) who observed the long-term upturn and downturn in the American telecommunications industry and concluded that that volatility and cyclicality will be an inherent part of the telecommunication sector in the future. First, in our thesis we explore the cyclical behavior of Taiwan telecommunications industry. As the turning point of the telecommunications industry may be obscure, we adopt a Markov Regime-Switching model with two regimes representing contraction and expansion. This nonlinear, two states, regime-switching model shows that Taiwan telecommunications industry has suffered from the cyclic fluctuation since the liberalization had been followed out. We focus on the mobile phone industry thereafter in this study. Since three telecommunication-related laws passed in 1996, the mobile phone industry is the first industry implemented the liberalization policy. In the process of the mobile phone industry's evolution, the carriers in this industry all experience the rapid growth in the mobile phone penetration rate and the fierce competition. Hence, to identify the main explanatory factors of the mobile phone industry fluctuation and cycles we introduce an 11-variable vector autoregressive (VAR) model. The empirical results confirm that the mobile phone industry' output can be influenced by five factors mainly including the macroeconomic status, demand, network effect, relative equipment import price, and output price, and furthermore, the impetus of the liberalization policy and the progress of the technology also play an important role beyond the five main factors in terms of the separate carriers' analysis.