CodingTrain / CodingTrain/Suggestion-Box

Evolution time

Abierto
#481 0 comentarios 0 reacciones 0 asignados Ver en GitHub
Lenguaje dominante
Sin datos de lenguaje
Estrellas
570
Forks
85
Métricas de merge de PR
Sin PR fusionados en 30 d

Descripción

I have an idea of decreasing evolution time by shifting gene index for fitness calculation.

Suppose we want to generate "To be or not to be". We compare every single letter to the result every time we want to calculate fitness. Code below increases comparing gene index starting point by 1 every 10th generation is made. Every time fitness calculation is active, comparing gene index will increase by 3.

The question arises. How can we be sure that every 3rd gene with starting point of 1 and 2 is "correct" after 20 generations? Maybe new way of reproduction or just adding an inspection array in fitness calculation method?

public class Evolution {

private final int TIME_INTERVAL = 10;
private final int GENE_SHIFT = 3;

private int geneStart = 0;

public Evolution(DNA resultDNA) {
startEvolution(resultDNA);
}

public void startEvolution(DNA resultDNA) {
initializePopulation();

int counter = 0;

while(true) {
counter++;

if(counter > TIME_INTERVAL) {
geneStart++;

if(geneStart > GENE_SHIFT - 1) {
geneStart = 0;
}

counter = 0;
}

calculateFitness(resultDNA.getGenes());
doSelection();
doReproduction();

if(evaluationOK()) {
break;
}
}
}

private void calculateFitness(Gene[] resultGenes) {
for(DNA dna : population) {
Gene[] genes = dna.getGenes();

dna.setFitness(0);

int i = geneStart;

while(i < genes.length) {
if(genes[i] == resultGenes[i]) {
dna.increaseFitness();
}

i += GENE_SHIFT;
}
}
}

private void doSelection() {
//...
}

private void doReproduction() {
//...
}
}

Guía de contribución

No hay ninguna guía de contribución indexada para este repositorio

Línea de trabajo

No repository file, test, or entry point is named. Start by locating any existing evolution or fitness-calculation implementation related to this proposal, then review whether the gene-shifting approach preserves fitness coverage across generations. Done is not defined by the issue; a contributor would need an agreed algorithm and validation criteria first.

Escrito por el modelo de indexación a partir del texto del issue.

Evaluación

Stack tecnológico
java
Área
machine-learning
Tipo de issue
Nueva funcionalidad
Dificultad
5/5
Tiempo estimado
Más de una semana
Estado de actividad
Estancado
Claridad
Necesita aclaración
Aptitud para principiantes
20/100

Recibe los nuevos issues en tu correo

Un resumen breve de issues de GitHub para principiantes.