Skip to content

features

Manlio Morini edited this page May 31, 2024 · 15 revisions

General features

  • GP / GA / Differential Evolution support.
  • Native support for concurrency.
  • Essentially parameterless (very very few mandatory parameters).
  • User programmable instruction set / termination criteria.
  • Loading and saving of populations and genetic programs (checkpoint and restoring mechanism).
  • Constraint handling.
  • CSV / Weka XRFF dataset file format compatibility.

Algorithms

  • Advanced linear genetic programming representation [1][2].
  • ALPS (Age-Layered Population Structure) [9].
  • Brood recombination [6][7]
  • Cache for improving performance.
  • Demetic grouping - Trivial Geography in Genetic Programming [3].
  • Differential evolution [4][5][8].
  • Gaussian Distribution for multi-class classification [13].
  • Self-adaptive crossover operator [14].
  • Steady state kernel.
  • STGP (Strongly Typed Genetic Programming) [10].
  • Team support [16].

Data

  • Holdout validation.
  • Stratified sampling.

Source

  • Completely ISO C++20 compliant.
  • Structured object oriented architecture.
  • Strict const-correctness.
  • Limited use of 3rd party libraries.
  • Comprehensive, searchable API documentation generated from the source (Doxygen).