Island model genetic programming based on frequent trees

Ono, Kenji; Hanada, Yoshiko; Kumano, Masahito; Kimura, Masahiro

doi:10.1109/cec.2013.6557933

Cited by 11 publications

(4 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We have proposed a migration strategy for the island model GP based on this hypothesis [20], and demonstrated performance improvement using benchmark problems widely adopted in the literature. We assume that the frequent trees, namely, the relatively small semi-structured subtrees frequently appearing among elite individuals in a pop ulation, are candidates for good pieces, where elite individuals have lower objective function values compared with the other individuals in a population.…”

Section: Proposed Assembling Bloat Control Strategiesmentioning

confidence: 99%

Assembling bloat control strategies in genetic programming for image noise reduction

Ono

Hanada

2014

2014 14th International Conference on Intelligent Systems Design and Applications

Self Cite

View full text Add to dashboard Cite

We address the problem of controlling bloat in genetic programming(GP) for image noise reduction. One of the most basic nonlinear filters for image noise reduction is the stack filter, and GP is suitable for estimating the min-max function used for a stack filter. However, bloat often occurs when the min-max function is estimated with GP. In order to enhance image noise reduction with GP, we extend the size-fair model GP, and propose a novel bloat control method based on tree size and frequent trees for image noise reduction, where the frequent trees are the relatively small subtrees appearing frequently among the population. By using texture images with impulse noise, we demonstrate that the size-fair model can achieve bloat control, and performance improvement can be achieved through bloat control based on tree size and frequent trees.Further, we demonstrate that the proposed method outperforms a typical image noise reduction method.

show abstract

Section: Proposed Assembling Bloat Control Strategiesmentioning

confidence: 99%

Assembling bloat control strategies in genetic programming for image noise reduction

Ono

Hanada

2014

2014 14th International Conference on Intelligent Systems Design and Applications

Self Cite

View full text Add to dashboard Cite

show abstract

“…In order to improve the performance of GP, various methods have been proposed: the method generating individuals in the next generation by joining fragments of the tree structure that randomly been sampled from several parent individuals [5], the method extracting useful tree structures from individuals called frequent trees [6], [7] that are subtrees that frequently appear in the population, the island model that combines those frequent trees [8], the method using the Semantic Aware Crossover (SAC) [9] that uses the similarity of subtrees to avoid destructive of tree structures, and the method in which semantics are used for select operation to keep diversity [10].…”

Section: Introductionmentioning

confidence: 99%

Extension of Genetic Programming with Multiple Trees for Agent Learning

Igarashi¹,

Takahashi²,

Inaba³

2016

JCP

View full text Add to dashboard Cite

This paper proposes an extension of genetic programming (GP) with multiple trees. In order to improve the performance, GP with control node (GPCN) and its three kinds of modification have been proposed. In GPCN, an individual consists of several trees which have the number P of executions. In previous work, the two kinds of modification, the conditional probability and the cross-cultural island model are employed. This paper proposes two methods: the new island model that combines the conditional probability with two islands in the cross-cultural island model and a method exchanges multiple trees in an individual in a suitable order. Experiments are conducted to show the performance in the garbage collection problem and the Santa Fe Trail problem.

show abstract

“…As a similar and more general idea, frequent trees, i.e., subtrees that frequently appear in the population, have been proposed. Chunks of strongly related nodes are regarded as frequent subtrees (Ono et al, 2012;Ono et al, 2013). In GP CN_CP , individuals in the next generation are generated using either genetic operations or conditional probability tables, where the conditional probability tables are updated using individuals with high fitness values.…”

Section: Introductionmentioning

confidence: 99%

“…We apply traditional GNP CN and GP CN and our methods, GP CN_CP , GP CN_IL , GP CN_CP (e), and GP CN_IL (e) to a garbage collection problem and Santa Fe Trail problem to compare the performance. We use these problems because the garbage collection problem and the Santa Fe Trail are used to show the ability of GNP and GP (Koza, 1992;Mesot et al, 2002;Eto et al, 2007;Ono et al, 2013;Iwashita and Iba, 2002). Although the symbolic regression problem exists as another type of benchmark problem for GP and GNP, we chose the garbage collection problem and the Santa Fe Trail problem because the objective of this paper is to obtain rules for agent actions.…”

Section: Introductionmentioning

confidence: 99%

Experiments Assessing Learning of Agent Behavior using Genetic Programming with Multiple Trees

Igarashi

Takahashi

Inaba

2014

Proceedings of the 6th International Conference on Agents and Artificial Intelligence

View full text Add to dashboard Cite

In this paper, experiments to assess agent behavior learning are conducted to demonstrate the performance of genetic programming (GP) with multiple trees. Using the methods, each has a chromosome representing agent behavior as several trees. We have proposed two variants using the conditional probability and the island model to improve the methods' performance. In GP using the conditional probability, individuals with high fitness values are used to produce conditional probability tables to generate individuals in the next generation. In GP using the island model, the population is divided into two islands of individuals: one island maintains diversity of individuals. The other emphasizes the accuracy of the solution. Moreover, this paper improves methods to seek the optimal number of executions of each tree in an individual. Those methods are applied to a garbage collection problem and a Santa Fe Trail problem. They are compared with traditional GP, GP with control nodes, and genetic network programming (GNP) with control nodes. Experimental results show that our methods are effective for improving the fitness.

show abstract

Island model genetic programming based on frequent trees

Cited by 11 publications

References 16 publications

Assembling bloat control strategies in genetic programming for image noise reduction

Assembling bloat control strategies in genetic programming for image noise reduction

Extension of Genetic Programming with Multiple Trees for Agent Learning

Experiments Assessing Learning of Agent Behavior using Genetic Programming with Multiple Trees

Contact Info

Product

Resources

About