B. Raphael

A direct stochastic algorithm for global search

This paper presents a new algorithm called probabilistic global search lausanne (PGSL). PGSL is founded on the assumption that optimal solutions can be identified through focusing search around sets of good solutions. Tests on benchmark problems having multi-parameter non-linear objective functions revealed that PGSL performs better than genetic algorithms and advanced algorithms for simulated annealing in 19 out of 23 cases studied. Furthermore as problem sizes increase, PGSL performs increasingly better than these other approaches. Empirical evidence of the convergence of PGSL is provided through its application to Lennard-Jones cluster optimisation problem. Finally, PGSL has already proved to be valuable for engineering tasks in areas of design, diagnosis and control.

Data mining techniques for improving the reliability of system identification

A system identification methodology that makes use of data mining techniques to improve the reliability of identification is presented in this paper. An important aspect of the methodology is the generation of a population of candidate models. Indications of the reliability of system identification are obtained through an examination of the characteristics of the population. Data mining techniques bring out model characteristics that are important. The methodology has been applied to several engineering systems.

A Bounded Index for Cluster Validity

Clustering is one of the most well known types of unsupervised learning. Evaluating the quality of results and determining the number of clusters in data is an important issue. Most current validity indices only cover a subset of important aspects of clusters. Moreover, these indices are relevant only for data sets containing at least two clusters. In this paper, a new bounded index for cluster validity, called the score function (SF), is introduced. The score function is based on standard cluster properties. Several artificial and real-life data sets are used to evaluate the performance of the score function. The score function is tested against four existing validity indices. The index proposed in this paper is found to be always as good or better than these indices in the case of hyperspheroidal clusters. It is shown to work well on multi-dimensional data sets and is able to accommodate unique and sub-cluster cases.

Generalized phase-shifting interferometry by use of a direct stochastic algorithm for global search

A new phase-shifting interferometric technique that uses an unknown phase step is described in which the phase step is determined by use of an algorithm called Probabilistic Global Search Lausanne (PGSL). One of the main sources of error in phase stepping is piezoelectric device (PZT) nonlinearity. The PGSL algorithm identifies the characteristics of the response of the PZT to the applied voltage through matching predicted and measured responses. The unknown phase step is also calculated with 0.097% error. This approach overcomes the limitations of existing techniques to determine unknown phase steps. Linear regression is subsequently applied for interference phase determination.

Lowering Costs of Timber Shear-Wall Design using Global Search

Abstract.This paper describes the application of a new global search technique, PGSL, that reduces structural costs for multi-storey timber housing. This technique is integrated within a computer aided design system that calculates production costs using a detailed product model. The cost calculation module was calibrated in collaboration with an industrial partner, and was tested on a full-scale completed project called Winter City 2000 in Sweden. It is found that the production cost for the shear-walls in the test case can be reduced by up to 10% due to an optimally directed search for structural designs. By using an explicit representation of the structural components in the shear-wall, it is also possible to increase the reliability of cost calculations.

Engineering Applications of a Direct Search Algorithm, PGSL

Note: CDROM Reference IMAC-CONF-2005-034 Record created on 2007-08-14, modified on 2016-08-08

Phase-shifting interferometry by a covariance-based method.

A novel generalized approach to phase-shifting interferometry in which phase distribution in an interferogram is evaluated in the presence of nonsinusoidal waveforms and piezoactuator device miscalibration is proposed. The approach is based on the underlying rotational invariance of signal subspaces spanned by two temporally displaced data sets. The advantage of the proposed method lies in its ability to identify arbitrary phase-step values pixelwise from an interference signal buried in noise. The robustness of the proposed method is investigated by addition of white Gaussian noise during the simulations.

Supporting engineers during system identification

Note: CDROM Reference IMAC-CONF-2005-035 Record created on 2007-08-14, modified on 2016-08-08

Creation of flexible graphical user interfaces through model composition

Nearly all software products have rigid and predefined interfaces. Users are usually unable to modify or customize features beyond cosmetic aspects. Interface adaptability is important because aspects such as user preferences and task sequences vary widely in engineering, even within specialized domains. A methodology for the creation of adaptable user interfaces using model composition is presented in this paper. User interfaces are generated dynamically through the composition of model fragments that are stored in a fragment library. When fragments are linked to models of physical behavior, interface model composition applications are likely to be easier to extend and maintain than traditional graphical user interfaces. A prototype system within the domain of bridge diagnosis illustrates the potential for practical applications.

Introduction of stochastic methods to phase-shifting interferometry

A new phase-shifting interferometric technique using an arbitrary phase step is described in which the step is determined using an algorithm called probabilistic global search Lausanne. The algorithm characterizes the piezo-actuator nonlinearity and precisely estimates the vibration induced during piezoelectric transducer translation. The unknown linearly varying phase step is calculated with around 0.082% error. A least-squares fit is subsequently applied for the measurement of interference phase once the exact phase step is known.