Genetic Programming Bloat without Semantics

@inproceedings{Langdon2000GeneticPB,
  title={Genetic Programming Bloat without Semantics},
  author={William B. Langdon and W. Banzhaf},
  booktitle={PPSN},
  year={2000}
}
To investigate the fundamental causes of bloat, six artificial random binary tree search spaces are presented. Fitness is given by program syntax (the genetic programming genotype). GP populations are evolved on both random problems and problems with "building blocks". These are compared to problems with explicit ineffective code (introns, junk code, inviable code). Our results suggest the entropy random walk explanation of bloat remains viable. The hard building block problem might be used in… 
Semantic and structural analysis of genetic programming
TLDR
An analysis of the process of initialisation using four novel algorithms to empirically evaluate specific traits of starting populations of programs reveals some interesting effects of evolution on program structure as well as offering evidence to support the success of the specialist operators.
Operator equalisation for bloat free genetic programming and a survey of bloat control methods
TLDR
This paper presents a comprehensive survey and taxonomy of many of the bloat control methods published in the literature through the years, and examines the evolutionary dynamics of OpEq and its potential to be extended and integrated into different elements of the evolutionary process.
Uniform Subtree Mutation
TLDR
Tests show that genetic programming using pure USM reduces evolved tree sizes dramatically, compared to crossover, but does impact solution quality somewhat, and in some cases, a combination of USM and crossover yielded both smaller trees and superior performance, as measured both by size effort and traditional metrics.
Improved Genetic Programming Based on Lineage Information
  • Hongxin Dong, Jia Chen
  • Computer Science
    2009 International Conference on Management and Service Science
  • 2009
TLDR
This paper introduces lineage relationship of chromosome in GP and proposes an improved lineage-based genetic programming algorithm, ILBGP: use of lineage information of several ancestors, at the same time only retains one chromosome with the same fitness randomly.
General Schema Theory for Genetic Programming with Subtree-Swapping Crossover: Part I
TLDR
A general schema theory for genetic programming (GP) with subtree-swapping crossover is introduced, based on a Cartesian node reference system which makes it possible to describe programs as functions over the space N2 and allows one to model the process of selection of the crossover points of subtree -swapping crossovers as a probability distribution over N4.
Dynamic limits for bloat control in genetic programming and a review of past and current bloat theories
TLDR
Dynamic Limits is introduced, a new approach to bloat control that implements a dynamic limit that can be raised or lowered, depending on the best solution found so far, and can be applied either to the depth or size of the programs being evolved.
A Simple but Theoretically-Motivated Method to Control Bloat in Genetic Programming
TLDR
This paper presents a simple method to control bloat based on the idea of strategically and dynamically creating fitness "holes" in the fitness landscape which repel the population by zeroing the fitness of a certain proportion of the offspring that have above average length.
A Schema Theory Analysis of the Evolution of Size in Genetic Programming with Linear Representations
TLDR
The schema theory presented in [20] is used to better understand the changes in size distribution when using GP with standard crossover and linear structures and implies the surprising result that a single program glitch in an otherwise flat fitness landscape is sufficient to drive the average program size of an infinite population.
Controlling Bloat in Genetic Programming for Sloving Wall Following Problem
TLDR
This paper review, evaluate, implementation and comparison of these methods in wall following problem and the most appropriate method for solving bloat problem is proposed.
Evolution of Type-correct Program Heuristics for Multi-Objective Combinatorial optimization problems Using Strongly Typed Genetic Programming
TLDR
A Type-basedGenetic Programming framework is developed to evolve a set of heuristics for the MOCO problems using minimal problem knowledge and the results obtained are found to be better than the un-typed implementation of GP.
...
1
2
3
...

References

SHOWING 1-10 OF 42 REFERENCES
Evolving computer programs without subtree crossover
TLDR
The evidence indicates that the technique is not only viable but is indeed capable of evolving good computer programs, and the results compare well with other evolutionary methods that rely on crossover to solve the same problems.
Genetic Programming Bloat with Dynamic Fitness
TLDR
Genetic programming, when evolving artificial ant control programs, is surprisingly little effected by large penalties and program growth is observed in all the authors' experiments.
Size Fair and Homologous Tree Genetic Programming Crossovers
Size fair and homologous crossover genetic operators for tree based genetic programming are described and tested. Both produce considerably reduced increases in program size and no detrimental eeect
Fitness Causes Bloat
The problem of evolving an artificial ant to follow the Santa Fe trail is used to study the well known genetic programming feature of growth in solution length. Known variously as “bloat”, “fluff”
Code growth in genetic programming
TLDR
It is found that without a constraint mechanism the programs will grow indefinitely regardless of whether or not the growth acts to improve the programs' solutions.
Data structures and genetic programming
TLDR
Proving that abstract data types (stacks, queues and lists) can be evolved using genetic programming and demonstrating that Genetic Algorithms can find low cost viable solutions to such problems demonstrate that data abstraction can be beneficial to automatic program generation via artificial evolution.
Analysis of Complexity Drift in Genetic
TLDR
The paper analyzes the in-uence of parsimony pressure on selection and growth of structures andenes a particular property of GP representations, called rooted tree-schema, that sheds light on the role of variable complexity of evolved structures.
The evolution of size and shape
TLDR
It is shown bloat in common operators is primarily due to the exponential shape of the underlying search space, and new operators with considerably reduced bloating characteristics are demonstrated, and the simple random walk entropy increasing model is able to predict the shape of evolved programs.
The evolution of size in variable length representations
  • W. Langdon
  • Biology
    1998 IEEE International Conference on Evolutionary Computation Proceedings. IEEE World Congress on Computational Intelligence (Cat. No.98TH8360)
  • 1998
TLDR
There are two causes of bloat: search operators with no length bias tend to sample bigger trees, and competition within populations favours longer programs, as they can usually reproduce more accurately.
Evolving Compact Solutions in Genetic Programming: A Case Study
TLDR
The experiments show that GP is an elegant and useful tool to derive complex functional dependencies on numerical data and several suggested methods are compared that are intended to improve GP performance and the readability of solutions by taking into account introns or redundancy that occurs in the trees and keeping the size of the trees small.
...
1
2
3
4
5
...