No description available.

An artificial neural network was used to optimize the release of salbutamol sulfate from hydrophilic matrix formulations. Model formulations to be used for training, testing and validating the neural network were manufactured with the aid of a central composite design with varying the levels of Methocel® K100M, xanthan gum, Carbopol® 974P and Surelease® as the input factors. In vitro dissolution time profiles at six different sampling times were used as target data in training the neural network for formulation optimization. A multi layer perceptron with one hidden layer was constructed using Matlab®, and the number of nodes in the hidden layer was optimized by trial and error to develop a model with the best predictive ability. The results revealed that a neural network with nine nodes was optimal for developing and optimizing formulations. Simulations undertaken with the training data revealed that the constructed model was useable. The optimized neural network was used for optimization of formulation with desirable release characteristics and the results indicated that there was agreement between the predicted formulation and the manufactured formulation. This work illustrates the possible utility of artificial neural networks for the optimization of pharmaceutical formulations with desirable performance characteristics.

M.Ing. / This dissertation discusses the results of a literature survey into the theoretical aspects and development of recurrent neural networks. In particular, the various architectures and training algorithms developed for recurrent networks are discussed. The various characteristics of importance for the efficient implementation of recurrent neural networks to model dynamical nonlinear processes have also been investigated and are discussed. Process control has been identified as a field of application where recurrent networks may play an important role. The model based adaptive control strategy is briefly introduced and the application of recurrent networks to both the direct- and the indirect adaptive control strategy highlighted. In conclusion, the important areas of future research for the successful implementation of recurrent networks in adaptive nonlinear control are identified

No description available.

A new method of realizing two-element-kind driving-point Impedances is given and illustrated by examples. In this method, networks of any desired topology and having a minimum of elements are utilized. A transformation to normal coordinates forms the basis of the method and, in order to determine network element values, evaluation of the associated transformation matrix is necessary. This matrix is found by formulating and solving a set of multivariable polynomial equations of second degree. The solution to this set of polynomial equations is obtained by a numerical perturbation procedure. To initiate the procedure, a set of element values is chosen, and the network of specified topology is analysed. The corresponding transformation matrix and driving-point impedance are determined from this analysis. The impedance parameters are then perturbed by small amounts in the direction of the specified ones, and the resulting changes in the transformation matrix are calculated. The process is continued until the transformation matrix corresponding to the specified impedance is obtained. A detailed description of the computer program written to carry out the above procedure is Included. A large number of examples of various complexities, including some canonic structures, have been realized by the method. Examples show the superiority of the numerical method to conventional procedures for solving multivariable nonlinear equations. In particular, the choice of the initial set of element values is not required to be close to the final set to achieve convergence to a solution. Some restrictions on the realizability of irreducible complementary tree structures are reported. It is shown that the specification parameters may have local extrema at a point where the Jacobian of the system of polynomial equations vanishes. Examples which support these results are given. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

A method of performance improvement of token ring networks is presented, based on the use of stations with two latency states. Station latency is defined as the time delay introduced in passing data through a station. Most token ring protocol standards (e.g. IEEE 802.5 or ANSI X3T9.5) require incoming data to be decoded and encoded in the station before transmission onto the ring. These encoding and decoding operations add significantly to the station latency. The bypassing of the encoding and decoding steps is proposed, which improves the mean message waiting time. A detailed evaluation and comparison of the networks is based on both analytical and simulation results. The performance of identical stations and symmetric traffic is obtained analytically. A discrete event simulation model for a token ring network is written in GPSS for general traffic. Results show a significant reduction in mean waiting time for the dual latency ring, with performance approaching or exceeding that of gated and exhaustive service, for certain ranges of network utilization. / Applied Science, Faculty of / Electrical and Computer Engineering, Department of / Graduate

Thesis (M. Phil.)--University of Hong Kong, 2006. / Title proper from title frame. Also available in printed format.

Thesis (M. Sc. (Computer Science))--Rhodes University, 2003.

Thesis (M.Phil.)--Hong Kong University of Science and Technology, 2008. / Includes bibliographical references (leaves 66-68). Also available in electronic version.

Thesis (doctoral)--Lund University, 1997. / Added t.p. with thesis statement inserted.