Probabilistic Model-Based Multistep Crossover Considering Dependency Between Nodes in Tree Optimization

Matsumura, Kohei; Hanada, Yoshiko; Ono, Kenji

doi:10.1007/978-3-319-62048-0_13

Cited by 1 publication

(1 citation statement)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In order to improve the performance of GP, various methods have been proposed: the method using the Semantic Aware Crossover (SAC) that uses the similarity of subtrees to avoid destructing of tree structures 3 , the method in which semantics is used for selecting operation to keep diversity 4 , and a mutation of subtrees and a crossover using semantic backpropagation 5 . In order to prevent the bloat phenomenon, GP removing a crossover operator and using multi-objective methods that add a distance between individuals to the objective function 6 , and multi-step search crossover based on neighbor search considering the dependency relationship of not only parent nodes but also child nodes 7 are proposed.…”

Section: Introductionmentioning

confidence: 99%

Obtaining the Number of Tree in the Evolution for Genetic Programming with Multiple Trees

Igarashi¹,

Takahashi²,

Inaba³

2017

Proceedings of the 5th IIAE International Conference on Industrial Application Engineering 2017

View full text Add to dashboard Cite

This paper proposes a method for automatic determination of the number of trees in genetic programming with control nodes (GP CN ). GP CN is genetic programing with multiple trees. An individual in GP CN comprises multiple trees, and each tree has a number P that indicates the number of repetitive actions based on the tree. In previous work, a method for inheriting the number P has been proposed to determine the number P in evolution. In this work, a method to determine the number of trees M in evolution is proposed. First, the fitness is changed to an advantageous fitness for individuals with small M. Second, a semantics for an agent's problem is proposed. Third mutation of the individual is proposed to generate individuals with various M for searching a suitable M.

show abstract

Section: Introductionmentioning

confidence: 99%