The manifold boundary approximation method, MBAM is a new technique used in approximating systems of equations using parameter reduction. This method and other approximation methods are introduced and described. Several current issues in performing MBAM are discussed in further detail. These issues significantly slow down the process of MBAM and create a barrier of entry for those wishing to use the method without a strong background in mathematics. A solution is proposed to automatically reparameterize models and evaluate specific types of variables approaching limits -- significantly speeding up the process of MBAM. An implementation of the solution is discussed.

The work of this thesis is concerned with the aggregation and advancement of modelling practise as used within modern day product development and optimisation environments making use of Model Based Design (¿MBD¿) and similar procedures. A review of model development and use forms the foundation of the work, with the findings being aggregated into two unique approaches for rapid model development and reconfiguration; the Plug-and-Simulate (¿PaS¿) approach and the Paradigm for Large Model Creation (¿PLMC¿); each shown to posses its own advantages. To support the MBD process a model optimisation algorithm that seeks to eliminate parameters that are of little or no significance to a simulation is developed. Eliminations are made on the basis of an energy analysis which determines the activity of a number of energy elements. Low activity elements are said to be of less significance to the global dynamics of a model and thus become targets for elimination. A model configuration tool is presented that brings together the PLMC and parameter elimination algorithm. The tool is shown to be useful for rapid configuration and reconfiguration of models and is capable of automatically running the optimisation algorithms thus producing a simulation model that is parametrically and computationally optimised. The response of the plug-and-simulate drivetrain submodels, assembled to represent a front wheel drive drivetrain, is examined. The resulting model is subjected to a torque step-input and an empirically obtained torque curve that characterises the input to a drivetrain undergoing steady acceleration. The model displays the expected response in both its full parameter and parameter reduced versions with simulation efficiency gains observed in the parameter reduced version. / EPSRC

The work of this thesis is concerned with the aggregation and advancement of modelling practise as used within modern day product development and optimisation environments making use of Model Based Design ('MBD') and similar procedures. A review of model development and use forms the foundation of the work, with the findings being aggregated into two unique approaches for rapid model development and reconfiguration; the Plug-and-Simulate ('PaS') approach and the Paradigm for Large Model Creation ('PLMC'); each shown to posses its own advantages. To support the MBD process a model optimisation algorithm that seeks to eliminate parameters that are of little or no significance to a simulation is developed. Eliminations are made on the basis of an energy analysis which determines the activity of a number of energy elements. Low activity elements are said to be of less significance to the global dynamics of a model and thus become targets for elimination. A model configuration tool is presented that brings together the PLMC and parameter elimination algorithm. The tool is shown to be useful for rapid configuration and reconfiguration of models and is capable of automatically running the optimisation algorithms thus producing a simulation model that is parametrically and computationally optimised. The response of the plug-and-simulate drivetrain submodels, assembled to represent a front wheel drive drivetrain, is examined. The resulting model is subjected to a torque step-input and an empirically obtained torque curve that characterises the input to a drivetrain undergoing steady acceleration. The model displays the expected response in both its full parameter and parameter reduced versions with simulation efficiency gains observed in the parameter reduced version.

In the current deep learning era, convolutional neural networks are commonly used as a backbone of systems that process images or videos. A lot of existing neural network architectures are however needlessly overparameterized and their performance can be closely matched by an alternative that uses much smaller amount of parameters. Our aim is to design a method that is able to find such alternative(s) for a given convolutional architecture. We propose a general scheme for architecture reduction and evaluate three algorithms that search for the op- timal smaller architecture. We do multiple experiments with ResNet and Wide ResNet architectures as the base using CIFAR-10 dataset. The best method is able to reduce the number of parameters by 75-85% without any loss in accuracy even in these already quite efficient architectures. 1

Práce se zabývá metodami pro testování lineárních analogových obvodů v kmitočtové oblasti. Cílem je navrhnout efektivní metody pro automatické generování testovacího plánu. Snížením počtu měření a výpočetní náročnosti lze výrazně snížit náklady za testování. Práce se zabývá multifrekveční parametrickou poruchovou analýzou, která byla plně implementována do programu Matlab. Vhodnou volbou testovacích kmitočtů lze potlačit chyby měření a chyby způsobené výrobními tolerancemi obvodových prvků. Navržené metody pro optimální volbu kmitočtů byly statisticky ověřeny metodou MonteCarlo. Pro zvýšení přesnosti a snížení výpočetní náročnosti poruchové analýzy byly vyvinuty postupy založené na metodě nejmenších čtverců a přibližné symbolické analýze.