Self-adaptation of parameters in a learning classifier system ensemble machine

Modeling of reliability characteristics typically assumes that components and systems fail if a certain individual damage level is exceeded. Every (mechanical) system damage increases irreversibly due to employed loading and (mechanical) stress, respectively. The main issue of damage estimation is adequate determination of the actual state-of-damage. Several mathematical modeling approaches are known in the literature, focusing on the task of how loading effects damage progression (e.g., Wöhler, 1870) for wear processes. Those models are only valid for specific loading conditions, a priori assumptions, set points, etc. This contribution proposes a general model, covering adequately the deterioration of a set of comparable systems under comparable loading. The main goal of this contribution is to derive the loading-damage connection directly from observation without assuming any damage models at the outset. Moreover, the connection is not investigated in detail (e.g., to examine the changes in material, etc.) but only approximated with a probabilistic approach. The idea is subdivided into two phases: A problem-specific relation between loading applied (to a structure, which contributes to the stress) and failure is derived from simulation, where a probabilistic approach only assumes a distribution function. Subsequently, an adequate general model is set up to describe deterioration progression. The scheme will be shown through simulation-based results and can be used for estimation of the remaining useful life and predictive maintenance/control.

show abstract

“…The calculation is very sensitive to wrongly chosen boundaries/assumptions, as mentioned before. As stated by Troć and Unold (2010), "[. .…”

Section: Classical Modelsmentioning

confidence: 99%

Adaptive modeling of reliability properties for control and supervision purposes

Dettmann

SéOuffker

2011

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

show abstract

“…EAs may be successfully combined with neural networks (Styrcz et al, 2011), reinforcement learning (Krawiec et al, 2011), ensemble machine learning methods (Troć and Unold, 2010) or heuristics which reduce the search space (Belter and Skrzypczyński, 2010). Hybrid approaches may also be utilised at various stages of EAs, including crossover (Jóźwiak and Postula, 2002).…”

Section: Crossover Operatorsmentioning

confidence: 99%

Customized crossover in evolutionary sets of safe ship trajectories

Szłapczynźki¹,

Szlapczynska²

2012

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

The paper presents selected aspects of evolutionary sets of safe ship trajectories-a method which applies evolutionary algorithms and some of the assumptions of game theory to solving ship encounter situations. For given positions and motion parameters of the ships, the method finds a near optimal set of safe trajectories of all ships involved in an encounter. The method works in real time and the solutions must be returned within one minute, which enforces speeding up the optimisation process. During the development of the method the authors tested various problem-dedicated crossover operators to obtain the best performance. The results of that research are given here. The paper includes a detailed description of these operators as well as statistical simulation results and examples of experiment results.

show abstract

“…The main idea of the learning procedure was based on the AdaSS algorithm , which uses an evolutionary approach (Ashlock, 2006;Troć and Unold, 2010), but we introduced several important changes such as the structure of the chromosome and improvement of some of the steps mainly associated with establishing mutation probabilities and protecting against overfitting.…”

Section: Training Algorithmmentioning

confidence: 99%

Combined classifier based on feature space partitioning

Woźniak¹,

Krawczyk²

2012

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

This paper presents a significant modification to the AdaSS (Adaptive Splitting and Selection) algorithm, which was developed several years ago. The method is based on the simultaneous partitioning of the feature space and an assignment of a compound classifier to each of the subsets. The original version of the algorithm uses a classifier committee and a majority voting rule to arrive at a decision. The proposed modification replaces the fairly simple fusion method with a combined classifier, which makes a decision based on a weighted combination of the discriminant functions of the individual classifiers selected for the committee. The weights mentioned above are dependent not only on the classifier identifier, but also on the class number. The proposed approach is based on the results of previous works, where it was proven that such a combined classifier method could achieve significantly better results than simple voting systems. The proposed modification was evaluated through computer experiments, carried out on diverse benchmark datasets. The results are very promising in that they show that, for most of the datasets, the proposed method outperforms similar techniques based on the clustering and selection approach.

show abstract

Self-adaptation of parameters in a learning classifier system ensemble machine

Cited by 13 publications

References 35 publications

Adaptive modeling of reliability properties for control and supervision purposes

Adaptive modeling of reliability properties for control and supervision purposes

Customized crossover in evolutionary sets of safe ship trajectories

Combined classifier based on feature space partitioning

Contact Info

Product

Resources

About