[bibshow file=my-publications.bib show_links=1 format=custom-ieee template=custom-bibshow highlight=”P. Sequeira”]
Genetica.NET
In general terms, GP allows a population of candidate programs to change over time by means of operators inspired from natural evolution such as selection, mutation and crossover. The evolutionary process is guided by a fitness function that assesses how fit a program is (usually its output) in regard to some external objective function.
Works that used this code: [bibcite key=sequeira2018cogsci]
Features
- Creation of programs as mathematical expressions
- Terminals: constant and variables
- Functions: arithmetic functions, sine, cosine, min, max, log, exponentiation and ‘if’ conditional operator
- Genetic operators
- Selection: tournament, roulette wheel (even and uneven selectors), stochastic
- Crossover: uniform, one-point, sub-tree, context-preserving, stochastic
- Mutation: point, sub-tree, hoist, shrink, swap, simplify, fitness simplify, stochastic
- Generation: full-depth, grow, stochastic
- Population class implementing a standard steady-state GP evolutionary procedure
- Rank (linear and non-linear) fitness functions
- Measure the similarity between two programs
- Similarity measures: value (according to the range of variables), primitive, leaf, sub-program, sub-combination, prefix and normal notation expression edit, tree edit, common region, average
- Conversion of programs to/from strings in normal and prefix notation
- Program simplification to remove redundancies and evolutionary noise
- Visual instruments (trees) to analyze the structure of sets of programs (e.g., a population):
- Information, symbol, ordered symbol, sub-program
Related Publications
[/bibshow]