Long-term evolution of genetic programming populations

@article{Langdon2017LongtermEO,
  title={Long-term evolution of genetic programming populations},
  author={William B. Langdon},
  journal={Proceedings of the Genetic and Evolutionary Computation Conference Companion},
  year={2017}
}
  • W. Langdon
  • Published 24 March 2017
  • Biology
  • Proceedings of the Genetic and Evolutionary Computation Conference Companion
Evolving binary mux-6 trees for up to 100 000 generations, during which some programs grow to more than a hundred million nodes, suggests the landscape which GP explores contains some very smooth regions. Although the GP population evolves under crossover, our unbounded GP appears not to evolve building blocks. We do see periods of tens even hundreds of generations where even although each member of the population occupies a different point in the genotypic search space, they are lie at exactly… 

Figures from this paper

Long-Term Evolution Experiment with Genetic Programming

It is suggested that deep expressions are resilient to learning as they disperse information, impeding evolvability, and the adaptation of highly nested organisms, and it is argued instead for open complexity.

Genetic Programming Convergence [Hot of the Press]

Both genotypic and phenotypic convergence in GP float-ing point continuous domain symbolic regression over thousands of generations are studied and near zero crossover disruption suggests improved GP systems.

Evolving open complexity

It is suggested that to evolve very large, very complex programs, it will be necessary to adopt an open architecture where most mutation sites are within 10--100 levels of the organism's environment.

Continuous Long-Term Evolution of Genetic Programming

This work evolves floating point Sextic polynomial populations of genetic programming binary trees for up to a million generations and uses both SIMD parallel AVX 512 bit instructions and 48 threads to yield performance of up to 149 billion GP operations per second.

Genetic programming convergence

  • W. Langdon
  • Computer Science
    Genet. Program. Evolvable Mach.
  • 2022
Despite large regions of zero variation, fitness continues to evolve and near zero crossover disruption suggests improved GP systems within existing memory use.

Deep Genetic Programming Trees Are Robust

  • W. Langdon
  • Computer Science
    ACM Transactions on Evolutionary Learning and Optimization
  • 2022
The genetic programming tree search space is sample and it is smooth, since many mutations on many test cases have little or no fitness impact, and deeply nested expressions are robust to crossover syntax changes, bugs, errors, run time glitches, perturbations, and so on.

Fitness First

  • W. Langdon
  • Computer Science
    Genetic and Evolutionary Computation
  • 2022
With side effect free terminals and functions it is possible to evaluate the fitness of genetic programming trees from their parents without creating them. This allows selection before forming the

Faster Genetic Programming GPquick via multicore and Advanced Vector Extensions

To support unbounded Long-Term Evolution Experiment LTEE GP, both SIMD parallel AVX 512 bit instructions and 48 threads are used to yield performance of up to 139 billion GP operations per second, 139 giga GPops, on a single Intel Xeon Gold 6126 2.60GHz server.

Incremental Evaluation in Genetic Programming

Sometimes GP evolves side effect free trees, which gives rise to functional resilience to source code changes, and GP convergence is viewed via information theory as evolving a smooth landscape and software plasticity.

Genetic programming is naturally suited to evolve bagging ensembles

It is shown that minor changes to fitness evaluation and selection are sufficient to make a simple and otherwise-traditional GP algorithm evolve ensembles efficiently, and this algorithm fares very well against state-of-the-art ensemble and non-ensemble GP algorithms.

References

SHOWING 1-10 OF 18 REFERENCES

Tweaking a tower of blocks leads to a TMBL: Pursuing long term fitness growth in program evolution

This work introduces Tweaking Mutation Behaviour Learning (TMBL), a form of evolutionary computation designed to reduce limits on long term fitness growth by encouraging tweaks: changes which affect behaviour without ruining the existing functionality.

A Field Guide to Genetic Programming

A unique overview of this exciting technique is written by three of the most active scientists in GP, which starts from an ooze of random computer programs, and progressively refines them through processes of mutation and sexual recombination until high-fitness solutions emerge.

Quadratic Bloat in Genetic Programming

Measurements indicate subtree crossover ceases to be disruptive with large programs and the population effectively converges (even though variety is near unity) and run time and memory O(no. generations1.0-2.0) are predicted, depending upon implementation.

Sub-machine-code genetic programming

Over the years a number of ideas on how to improve GP performance have been proposed in the literature and some of these techniques are now used in many GP implementations.

A Schema Theory Analysis of the Evolution of Size in Genetic Programming with Linear Representations

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.

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

Foundations of Genetic Programming

This is one of the only books to provide a complete and coherent review of the theory of genetic programming (GP) and presents new exact schema analysis, showing that it applies to GP as well as to the simpler GAs.

On the Limiting Distribution of Program Sizes in Tree-Based Genetic Programming

We provide strong theoretical and experimental evidence that standard sub-tree crossover with uniform selection of crossover points pushes a population of a-ary GP trees towards a distribution of

Recombination, selection, and the genetic construction of computer programs

Analysis is performed which shows that Genetic Programming has a great deal in common with heuristic search and introduces a novel aspect to that method in the form of the recombination operator which generates successors by combining parts of favorable strategies.

Sustained fitness gains and variability in fitness trajectories in the long-term evolution experiment with Escherichia coli

The results imply that both adaptation and divergence can continue indefinitely— or at least for a long time—even in a constant environment.