Self adaptive island GA

Takashima, Eiichi; Murata, Yoshinori; Shibata, Naoki; Ito, Minoru

doi:10.1109/cec.2003.1299787

Cited by 8 publications

(6 citation statements)

References 8 publications

(7 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Some good examples of adaptive strategies for hybrid EAs are the fuzzy logic controller proposed in [9] to control the participation of different crossover operators or the work in [10] where the participation of several crossover operators is adjusted based on the progress introduced in the population by using each of them. Other algorithms propose island GAs where each island evolves a population by means of variation operators with different characteristics, trying to exploit exploration and exploitation mechanisms by controlling migratory processes [11], [12]. Finally, some studies propose the use of two different populations: one for the problem itself and another one for the set of operators that will be used [13].…”

Section: Preliminariesmentioning

confidence: 99%

Quality measures to adapt the participation in MOS

LaTorre

Peña

Muelas

et al. 2009

2009 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

Multiple Offspring Sampling (MOS) is a hybrid algorithm where different evolutionary approaches can coexist simultaneously. The algorithm dynamically evaluates the quality of the solutions produced by each of these algorithms (or techniques, as they are called within MOS) and adjusts their participation in the overall evolutionary process according to this quality value. In this paper we use two alternative measures to evaluate the quality of a reproductive technique and therefore perform the dynamic adjustment of the participation ratios. One of these measures considers the fitness values of the solutions, while the other one determines how difficult the problem is for an evolutionary approach. These two measures are tested and compared over four problems of different complexity and domain (three of them are continuous while the fourth one is discrete). Results show analogies and differences among the used measures and confirm that a good dynamic selection based on a quality measure can boost the performance of the hybrid algorithm.

show abstract

Section: Preliminariesmentioning

confidence: 99%

Quality measures to adapt the participation in MOS

LaTorre

Peña

Muelas

et al. 2009

2009 IEEE Congress on Evolutionary Computation

View full text Add to dashboard Cite

show abstract

“…Some good examples in this line are the fuzzy logic controller proposed in [11] to control the participation of different crossover operators or the experiments by Hong [13] where the participation of several crossover operators is adjusted based on the progress introduced into the population by using each of them. Other algorithms propose island GAs where each island evolves a population by means of recombination operators with different characteristics, trying to achieve a good trade-off between exploration and exploitation mechanisms by controlling migratory processes [30]. Finally, some studies propose the use of two different populations: one for the problem itself and another one for the set of operators that will be used [17].…”

Section: Hybrid Evolutionary Algorithmsmentioning

confidence: 99%

Learning hybridization strategies in evolutionary algorithms

LaTorre

Peña

Muelas

et al. 2010

IDA

View full text Add to dashboard Cite

Evolutionary Algorithms are powerful optimization techniques which have been applied to many different problems, from complex mathematical functions to real-world applications. Some studies report performance improvements through the combination of different evolutionary approaches within the same hybrid algorithm. However, the mechanisms used to control this combination of evolutionary approaches are not as satisfactory as would be desirable. In most cases, there is no feedback from the algorithm nor any regulatory component that modifies the participation of each evolutionary approach in the overall search process. In some cases, the algorithm makes use of some information for an on-line adaptation of the participation of each algorithm. In this paper, the use of Reinforcement Learning (RL) is proposed as a mechanism to control how the different evolutionary approaches contribute to the overall search process. In particular, three learning policies based on one of the state-of-the-art RL algorithms, Q-Learning, have been considered and used to control the participation of each algorithm by learning the best-response mixed strategy. To test this approach, a benchmark made up of six large-scale (500 dimensions) continuous optimization functions has been considered. The experimentation carried out has proved that RL control mechanisms successfully learn optimal patterns for the combination of Evolutionary Algorithms in most of the proposed functions, being able to improve the performance of both individual and non RL hybrid algorithms.

show abstract

“…Our approach is somewhat similar to that of (Opitz et al, 1996), where the ensemble of neural networks is optimized by means of a genetic algorithm. Because each XCS component applies the GA for rule discovery, our model is also similar to some parallel genetic algorithms, where an adaptation of parameters is made at the subpopulation level (Tongchim and Chongstitvatana, 2002;Takashima et al, 2003). The sensitiveness of XCS parameters and problems with their adaptation at the classifier level are the key motivation for our model.…”

Section: Self-adaptive Xcs-based Ensemble Machinementioning

confidence: 99%

“…It seems to be necessary because, in our previous model (Troć and Unold, 2008), the MEA is executed in some fixed number of iterations of components in which parameter values are optimized. Similarly, in related works (Tongchim and Chongstitvatana, 2002;Takashima et al, 2003), the step of meta genetic algorithm takes place in a predefined number of generations or fitness evaluations of individuals in subpopulations.…”

Section: Component Learning and Calling The Meta Evolutionary Algorithmmentioning

confidence: 99%

“…In this group of methods, which are in a majority derived from the meta genetic algorithm (meta-GA) (Grefenstette, 1986), the additional, distinguished evolutionary process is applied for searching the parameter space. In the versions proposed for parallel genetic algorithms (Tongchim and Chongstitvatana, 2002;Takashima et al, 2003), the model based on a population divided into several subpopulations evolving in parallel is applied. Each of them uses its own vector of parameter values.…”

Section: Troć and O Unoldmentioning

confidence: 99%

See 1 more Smart Citation

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

Troć¹,

Unold²

2010

International Journal of Applied Mathematics and Computer Science

View full text Add to dashboard Cite

Self-adaptation is a key feature of evolutionary algorithms (EAs). Although EAs have been used successfully to solve a wide variety of problems, the performance of this technique depends heavily on the selection of the EA parameters. Moreover, the process of setting such parameters is considered a time-consuming task. Several research works have tried to deal with this problem; however, the construction of algorithms letting the parameters adapt themselves to the problem is a critical and open problem of EAs. This work proposes a novel ensemble machine learning method that is able to learn rules, solve problems in a parallel way and adapt parameters used by its components. A self-adaptive ensemble machine consists of simultaneously working extended classifier systems (XCSs). The proposed ensemble machine may be treated as a meta classifier system. A new self-adaptive XCS-based ensemble machine was compared with two other XCSbased ensembles in relation to one-step binary problems: Multiplexer, One Counts, Hidden Parity, and randomly generated Boolean functions, in a noisy version as well. Results of the experiments have shown the ability of the model to adapt the mutation rate and the tournament size. The results are analyzed in detail.

show abstract

Self adaptive island GA

Cited by 8 publications

References 8 publications

Quality measures to adapt the participation in MOS

Quality measures to adapt the participation in MOS

Learning hybridization strategies in evolutionary algorithms

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

Contact Info

Product

Resources

About