One of the challenges of data mining is finding hyper-parameters for a learning algorithm that will produce the best model for a given dataset. Hyper-parameter optimization automates this process, but it can still take significant time. It has been found that hyperparameter optimization does not always result in induced models with significant improvement over default hyper-parameters, yet no systematic analysis of the role of hyper-parameter optimization in machine learning has been conducted. We propose the use of meta-learning to inform the decision to optimize hyper-parameters based on whether default hyper-parameter performance can be surpassed in a given amount of time. We will build a base of metaknowledge, through a series of experiments, to build predictive models that will assist in the decision process.

Many ecological populations can be interpreted as response surfaces; the spatial patterns of the population vary in response to changes in the spatial patterns of environmental explanatory variables. Collection of a probability sample from the population provides unbiased estimates of the population parameters using design based estimation. When information is available for the environmental explanatory variables, model based procedures are available that provide more precise estimates of population parameters in some cases. In practice, not all of these environmental explanatory variables will be known. When the spatial coordinates of the population units are available, a spatial model can be used as a surrogate for the unknown, spatially patterned explanatory variables. Design based and model based procedures will be compared for estimating parameters of the population of Acid Neutralizing Capacity (ANC) of lakes in the Adirondack Mountains in New York. Results from the analysis of this population will be used to elucidate some general principles for model based estimation of parameters of spatial populations. Results indicate that using model based estimates of population parameters provide more precise estimates than design based estimates in some cases. In addition, including spatial information as a surrogate for spatially patterned missing covariates improves the precision of the estimates in some cases, the degree to which depends upon the model chosen to represent the spatial pattern. When the probability sample is selected from the spatial population is a stratified sample, differences in stratum variances need to be accounted for when residual spatial covariance estimation is desired for the entire population. This can be accomplished by scaling residuals by their estimated stratum standard deviation functions, and calculating the residual covariance using these scaled residuals. Results here demonstrate that the form of scaling influences the estimated strength of the residual correlation and the estimated correlation range. / Graduation date: 1997

Fractional order calculus (FOC) has wide applications in modeling natural behavior of systems related to different areas of engineering including bioengineering, viscoelasticity, electronics, robotics, control theory and signal processing. This thesis aims at modeling a lossy transmission line using fractional order calculus and identifying its parameters. A lossy transmission line is considered where its behavior is modeled by a fractional order transfer function. A semi-infinite lossy transmission line is presented with its distributed parameters R, L, C and ordinary AC circuit theory is applied to find the partial differential equations. Furthermore, applying boundary conditions and the Laplace transformation a generalized fractional order transfer function of the lossy transmission line is obtained. A finite length lossy transmission line terminated with arbitrary load is also considered and its fractional order transfer function has been derived. Next, the frequency responses of lossy transmission lines from their fractional order transfer functions are also derived. Simulation results are presented to validate the frequency responses. Based on the simulation results it can be concluded that the derived fractional order transmission line model is capable of capturing the phenomenon of a distributed parameter transmission line. The achievement of modeling a highly accurate transmission line requires that a realistic account needs to be taken of its parameters. Therefore, a parameter identification technique to identify the parameters of the fractional order lossy transmission line is introduced. Finally, a few open problems are listed as the future research directions. / Controls

A networked control system (NCS) is a feedback control system that has its control loop physically connected via real-time communication networks. To meet the demands of `teleautomation', modularity, integrated diagnostics, quick maintenance and decentralization of control, NCSs have received remarkable attention worldwide during the past decade. Yet despite their distinct advantages, NCSs are suffering from network-induced constraints such as time delays and packet dropouts, which may degrade system performance. Therefore, the network-induced constraints should be incorporated into the control design and related studies.<p> For the problem of state estimation in a network environment, we present the strategy of simultaneous input and state estimation to compensate for the effects of unknown input missing. A sub-optimal algorithm is proposed, and the stability properties are proven by analyzing the solution of a Riccati-like equation.<p> Despite its importance, system identification in a network environment has been studied poorly before. To identify the parameters of a system in a network environment, we modify the classical Kalman filter to obtain an algorithm that is capable of handling missing output data caused by the network medium. Convergence properties of the algorithm are established under the stochastic framework.<p> We further develop an adaptive control scheme for networked systems. By employing the proposed output estimator and parameter estimator, the designed adaptive control can track the expected signal. Rigorous convergence analysis of the scheme is performed under the stochastic framework as well.

Aircraft with advanced wing geometry, like the flying wing or blended wing body configuration, seems to be the seed candidate of future aircraft. Compared with conventional aircraft, there are significant aerodynamic performance improvements because of its highly integrated wing and fuselage configuration. On the other hand, due to its tailless configuration, the stability characteristics are not as good as conventional aircraft. The research aims to compare the aerodynamic and stability characteristics of conventional, flying wing and blended wing body aircraft. Based on the same requirement—250 passenger capability and 7,500 nautical miles range, three different configurations—conventional, flying wing and blended wing body options were provided to make direct comparison. The research contains four parts. In the first part, the aerodynamic characteristics were compared using empirical equation ESDU datasheet and Vortex-Lattice Method based AVL software. In the second part, combined with the aerodynamic data and output mass data from other team member, the stability characteristics were analysed. The stability comparison contains longitudinal, lateral-directional static stability and dynamic stability. In the third part, several geometry parameters were varied to investigate the influence on the aerodynamic and stability characteristics of blended wing body configuration. In the last part, a special case has been explored in an attempt to improve the static stability by changing geometry parameters. The process shows that the design of blended wing body is really complex since the closely coupling of several parameters.

This thesis adresses the problem of estimating the parameters in a permanent magnet synchronous motor (PMSM). There is an uncertainty about the parameters, due to age and tolerances in the manufacturing process. Parameters such as the resistance and the current to torque factor Kt, changes with respect to temperature as well. The temperature in the motor varies in normal motor operation, due to variations in angular velocity and torques. Online estimation methods with the model reference adaptive systems technique (MRAS) and offline methods are presented. The estimation algorithms are validated in simulations with Matlab/Simulink and also evaluated with experimental data. Experiments were performed on a range of different motors, in realistic scenarios. Relevant factors such as the angular velocity of the rotor and the impact of the gravity force are investigated. The results show that it is possible to estimate the motor factor $K_t$, with an accuracy of two percentage from its reference value in normal industry conditions. The estimated value of the motor inductance is within 25 percentage of the calculated reference value. The resistance however is affected by the resistance in the cables from the motor to the measurement device. With the cable resistance included in the calculations, the estimate still often exceeds double the value of the reference value.

A networked control system (NCS) is a feedback control system that has its control loop physically connected via real-time communication networks. To meet the demands of `teleautomation', modularity, integrated diagnostics, quick maintenance and decentralization of control, NCSs have received remarkable attention worldwide during the past decade. Yet despite their distinct advantages, NCSs are suffering from network-induced constraints such as time delays and packet dropouts, which may degrade system performance. Therefore, the network-induced constraints should be incorporated into the control design and related studies.<p> For the problem of state estimation in a network environment, we present the strategy of simultaneous input and state estimation to compensate for the effects of unknown input missing. A sub-optimal algorithm is proposed, and the stability properties are proven by analyzing the solution of a Riccati-like equation.<p> Despite its importance, system identification in a network environment has been studied poorly before. To identify the parameters of a system in a network environment, we modify the classical Kalman filter to obtain an algorithm that is capable of handling missing output data caused by the network medium. Convergence properties of the algorithm are established under the stochastic framework.<p> We further develop an adaptive control scheme for networked systems. By employing the proposed output estimator and parameter estimator, the designed adaptive control can track the expected signal. Rigorous convergence analysis of the scheme is performed under the stochastic framework as well.

We revisited an empirical relationship between the integrated volume depolar- ization ratio, oacc, and the effective multiple scattering factor, -n, on the basis of Monte Carlo simulations of spaceborne lidar backscatter associated with homogeneous wa- ter clouds. The relationship is found to be sensitive to the extinction coefficient and to the particle size. The layer integrated attenuated backscatter is also obtained. Comparisons made between the simulations and statistics derived relationships of the layer integrated depolarization ratio, oacc, and the layer integrated attenuated backscatter, -n, based on the measurement by the Cloud-Aerosol Lidar and Infrared Pathfinder Satellite Observations (CALIPSO) satellite show that a cloud with a large effective size or a large extinction coefficient has a relatively large integrated backscatter and a cloud with a small effective size or a large extinction coefficient has a large integrated volume depolarization ratio. The present results also show that optically thin water clouds may not obey the empirical relationship derived by Y. X. Hu. and co-authors.

The current study has applied the Genetic Algorithm (GA) for the boundary parameters calibration in the hydrodynamic-based tidal model. The objective is to minimize the deviation between the estimated results acquired from the simulation model and the real tidal data along Taiwan coast. The manual trial-error has been widely used in the past, but such approach is inefficient due to the complexity posed by the tremendous amounts of parameters. Fortunately, with the modern computer capability, some automatic searching processes, in particular GA, can be implemented to handle the large data set and reduce the human subjectivity when conducting the calibration. Besides, owing to the efficient evolution procedures, GA can find better solutions in a shorter time compared to the manual approach. Based on the preliminary experiments of the current study, the integration of GA with the hydrodynamic-based tidal model can improve the accuracy of simulation.

Abstract Pool boiling process is frequently encountered in a number of engineering applications. It is difficult to exactly predict the heat transfer coefficient. This is because the boiling phenomenon is rather complex and influenced by many factors, such as surface condition, heater size, geometry, material, arrangement of heated rods, and refrigerants, etc. The key boiling parameters (bubble dynamics data) such as bubble departure diameter, frequency, velocity and nucleation site density will be varied in such different heated rod pitches resulting in the different effect of heat transfer. Furthermore, more fundamental of the physical phenomenon can be obtained. Pool boiling heat transfer of R-134a is investigated experimentally on twin tube arrangement. The tube pitch is 1.65 and 2.5. The surface condition was prepared with plasma spray coating. In addition, using the high-speed digital camera and LDV, the bubble diameter and dynamics of R-134a were measured while growing. The boiling curves in different twin-tube pitches were drawn and the influence of bubble velocity on heat transfer coefficients was also examined. Finally, to broaden our basic understanding of different arrangement of heated rods and heat transfer mechanisms, thermal design data of a flooded type evaporator of high performance as well as more and further physical insight of the above-stated nucleate boiling heat transfer can be acquired. The results would hopefully be helpful not only for the academia but also for the industry.