scholarly journals A Rigorous Evaluation of Crossover and Mutation in Genetic Programming

2009 ◽  
pp. 220-231 ◽  
Author(s):  
David R. White ◽  
Simon Poulding
Keyword(s):  
Genetic Programming ◽  
Rigorous Evaluation ◽  
Crossover And Mutation

scholarly journals Explorations in Parallel  Linear Genetic Programming

2021 ◽  
Author(s):  
◽  
Carlton Downey
Keyword(s):  
Genetic Programming ◽  
Convergence Time ◽  
Linear Genetic Programming ◽  
Cooperative Coevolution ◽  
Biological Inspiration ◽  
Linear Sequence ◽  
New Form ◽  
Program Output ◽  
Program Components ◽  
Crossover And Mutation

<p>Linear Genetic Programming (LGP) is a powerful problem-solving technique, but one with several significant weaknesses. LGP programs consist of a linear sequence of instructions, where each instruction may reuse previously computed results. This structure makes LGP programs compact and powerful, however it also introduces the problem of instruction dependencies. The notion of instruction dependencies expresses the concept that certain instructions rely on other instructions. Instruction dependencies are often disrupted during crossover or mutation when one or more instructions undergo modification. This disruption can cause disproportionately large changes in program output resulting in non-viable offspring and poor algorithm performance. Motivated by biological inspiration and the issue of code disruption, we develop a new form of LGP called Parallel LGP (PLGP). PLGP programs consist of n lists of instructions. These lists are executed in parallel, and the resulting vectors are summed to produce the overall program output. PLGP limits the disruptive effects of crossover and mutation, which allows PLGP to significantly outperform regular LGP. We examine the PLGP architecture and determine that large PLGP programs can be slow to converge. To improve the convergence time of large PLGP programs we develop a new form of PLGP called Cooperative Coevolution PLGP (CC PLGP). CC PLGP adapts the concept of cooperative coevolution to the PLGP architecture. CC PLGP optimizes all program components in parallel, allowing CC PLGP to converge significantly faster than conventional PLGP. We examine the CC PLGP architecture and determine that performance</p>

Some Improvements of Genetic Programming in Data Fitting

2011 ◽  
Vol 201-203 ◽  
pp. 2536-2539
Author(s):  
Hu Jie ◽  
Jia Quan Feng ◽  
Da Lin Chen
Keyword(s):  
Numerical Simulation ◽  
Genetic Programming ◽  
Data Fitting ◽  
Parse Tree ◽  
Tree Expansion ◽  
Improvement Measures ◽  
Crossover And Mutation

This paper proposed some improvement measures of Genetic Programming (GP) in data fitting, including developed new ways of crossover and mutation, improved the calculation efficiency greatly, and avoided the problem of parse tree expansion. The new adopted mutation method improved the problem of constant modification to some extent. Numerical simulation obtained a considerable good fitting and prediction precision.

An Intelligent Modeling Method Based on Genetic Programming and Genetic Algorithm

2008 ◽  
Vol 33-37 ◽  
pp. 795-800 ◽  
Author(s):  
Jie Hu ◽  
Xi Nong Zhang ◽  
Shi Lin Xie
Keyword(s):  
Genetic Algorithm ◽  
Genetic Programming ◽  
Approximate Model ◽  
Parse Tree ◽  
Experiment Data ◽  
Intelligent Modeling ◽  
Fitness Value ◽  
Strain Equation ◽  
Matlab Programming ◽  
Crossover And Mutation

This paper utilizes Genetic Programming(GP) and Genetic Algorithm(GA) to analyze experiment data. The purpose of this research is to establish a function model of the data. The core methodology of this research is using GP to get the approximate model first, and then optimizes the parameters and enhance the fitness value of the model by using GA. To validate this method, two examples are given: one is the reconstruction of permeability-strain equation of the rock in literature[1]; another example is the function search automatically of the wire cable isolator experiment data. In the process of programming of parse tree, this paper adopted a new way that different from three traditional methods, the parse tree is described by matrix of special size, more significantly, this new method makes the genetic operation of crossover and mutation intuitionstic, even the pellucid Matlab programming language could implement it.

scholarly journals Explorations in Parallel  Linear Genetic Programming

2021 ◽  
Author(s):  
◽  
Carlton Downey
Keyword(s):  
Genetic Programming ◽  
Convergence Time ◽  
Linear Genetic Programming ◽  
Cooperative Coevolution ◽  
Biological Inspiration ◽  
Linear Sequence ◽  
New Form ◽  
Program Output ◽  
Program Components ◽  
Crossover And Mutation

<p>Linear Genetic Programming (LGP) is a powerful problem-solving technique, but one with several significant weaknesses. LGP programs consist of a linear sequence of instructions, where each instruction may reuse previously computed results. This structure makes LGP programs compact and powerful, however it also introduces the problem of instruction dependencies. The notion of instruction dependencies expresses the concept that certain instructions rely on other instructions. Instruction dependencies are often disrupted during crossover or mutation when one or more instructions undergo modification. This disruption can cause disproportionately large changes in program output resulting in non-viable offspring and poor algorithm performance. Motivated by biological inspiration and the issue of code disruption, we develop a new form of LGP called Parallel LGP (PLGP). PLGP programs consist of n lists of instructions. These lists are executed in parallel, and the resulting vectors are summed to produce the overall program output. PLGP limits the disruptive effects of crossover and mutation, which allows PLGP to significantly outperform regular LGP. We examine the PLGP architecture and determine that large PLGP programs can be slow to converge. To improve the convergence time of large PLGP programs we develop a new form of PLGP called Cooperative Coevolution PLGP (CC PLGP). CC PLGP adapts the concept of cooperative coevolution to the PLGP architecture. CC PLGP optimizes all program components in parallel, allowing CC PLGP to converge significantly faster than conventional PLGP. We examine the CC PLGP architecture and determine that performance</p>

SEMANTIC SEARCH TECHNIQUES FOR LEARNING SMALLER BOOLEAN EXPRESSION TREES IN GENETIC PROGRAMMING

2014 ◽  
Vol 13 (03) ◽  
pp. 1450018
Author(s):  
NICHOLAS C. MILLER ◽  
PHILIP K. CHAN
Keyword(s):  
Genetic Programming ◽  
Semantic Search ◽  
Experimental Results ◽  
Boolean Expression ◽  
Mutation Operators ◽  
Recent Interest ◽  
Program Semantics ◽  
Large Program ◽  
Crossover And Mutation ◽  
Better Than

One sub-field of Genetic Programming (GP) which has gained recent interest is semantic GP, in which programs are evolved by manipulating program semantics instead of program syntax. This paper introduces a new semantic GP algorithm, called SGP+, which is an extension of an existing algorithm called SGP. New crossover and mutation operators are introduced which address two of the major limitations of SGP: large program trees and reduced accuracy on high-arity problems. Experimental results on "deceptive" Boolean problems show that programs created by the SGP+ are 3.8 times smaller while still maintaining accuracy as good as, or better than, SGP. Additionally, a statistically significant improvement in program accuracy is observed for several high-arity Boolean problems.

Crossover and mutation operators for grammar-guided genetic programming

2006 ◽  
Vol 11 (10) ◽  
pp. 943-955 ◽  
Author(s):  
Jorge Couchet ◽  
Daniel Manrique ◽  
Juan Ríos ◽  
Alfonso Rodríguez-Patón
Keyword(s):  
Genetic Programming ◽  
Mutation Operators ◽  
Crossover And Mutation

Defining the Competence of Psychologists with a View to Public Accountability *

1997 ◽  
Vol 2 (4) ◽  
pp. 293-300 ◽  
Author(s):  
Ype H. Poortinga ◽  
Ingrid Lunt
Keyword(s):  
Scientific Knowledge ◽  
Professional Training ◽  
Professional Competence ◽  
Codes Of Ethics ◽  
The Body ◽  
The European Union ◽  
Rigorous Evaluation ◽  
Judicial Proceedings ◽  
Psychological Knowledge

In national codes of ethics the practice of psychology is presented as rooted in scientific knowledge, professional skills, and experience. However, it is not self-evident that the body of scientific knowledge in psychology provides an adequate basis for current professional practice. Professional training and experience are seen as necessary for the application of psychological knowledge, but they appear insufficient to defend the soundness of one's practices when challenged in judicial proceedings of a kind that may be faced by psychologists in the European Union in the not too distant future. In seeking to define the basis for the professional competence of psychologists, this article recommends taking a position of modesty concerning the scope and effectiveness of psychological interventions. In many circumstances, psychologists can only provide partial advice, narrowing down the range of possible courses of action more by eliminating unpromising ones than by pointing out the most correct or most favorable one. By emphasizing rigorous evaluation, the profession should gain in accountability and, in the long term, in respectability.

Enhanced Artificial Intelligent Method for Representing the Photonic Crystal Fiber Profile Index

2018 ◽  
Vol 1 (1) ◽  
pp. 2-19
Author(s):  
Mahmood Sh. Majeed ◽  
Raid W. Daoud
Keyword(s):  
Genetic Algorithms ◽  
Photonic Crystal ◽  
Photonic Crystal Fiber ◽  
New Method ◽  
Proper Solution ◽  
Artificial Intelligent ◽  
Crystal Fiber ◽  
Fitness Value ◽  
First Pass ◽  
Crossover And Mutation

A new method proposed in this paper to compute the fitness in Genetic Algorithms (GAs). In this new method the number of regions, which assigned for the population, divides the time. The fitness computation here differ from the previous methods, by compute it for each portion of the population as first pass, then the second pass begin to compute the fitness for population that lye in the portion which have bigger fitness value. The crossover and mutation and other GAs operator will do its work only for biggest fitness portion of the population. In this method, we can get a suitable and accurate group of proper solution for indexed profile of the photonic crystal fiber (PCF).

Sign in / Sign up

Export Citation Format

Share Document