Genetic Algorithm Selection Methods (Genetic Operator)
- Source:
- See:
Example
Copy
let { architect, methods } = require("@liquid-carrot/carrot");
let myNetwork = new architect.Perceptron(1,1,1);
let myTrainingSet = [{ input:[0], output:[1]}, { input:[1], output:[0]}];
myNetwork.evolve(myTrainingSet, {
generations: 10,
selection: methods.selection.POWER // eg.
});
Members
(static, constant) FITNESS_PROPORTIONATE :object
Type:
- object
Example
Copy
let { architect, methods } = require("@liquid-carrot/carrot");
let myNetwork = new architect.Perceptron(1,1,1);
let myTrainingSet = [{ input:[0], output:[1]}, { input:[1], output:[0]}];
myNetwork.evolve(myTrainingSet, {
generations: 10,
selection: methods.selection.FITNESS_PROPORTIONATE // eg.
});
- Source:
- Default Value:
{ "name": "FITNESS_PROPORTIONATE" }
(static, constant) POWER :object
A random decimal value between 0 and 1 will be generated (e.g. 0.5) then this value will get an exponential value (power, default is 4). So 0.5**4 = 0.0625. This is converted into an index for the array of the current population, sorted from fittest to worst.
Type:
- object
Example
Copy
let { architect, methods } = require("@liquid-carrot/carrot");
let myNetwork = new architect.Perceptron(1,1,1);
let myTrainingSet = [{ input:[0], output:[1]}, { input:[1], output:[0]}];
myNetwork.evolve(myTrainingSet, {
generations: 10,
selection: methods.selection.POWER // eg.
});
- Source:
- Default Value:
{ "name": "POWER", "power": 4 }
(static, constant) TOURNAMENT :object
Type:
- object
Example
Copy
let { architect, methods } = require("@liquid-carrot/carrot");
let myNetwork = new architect.Perceptron(1,1,1);
let myTrainingSet = [{ input:[0], output:[1]}, { input:[1], output:[0]}];
myNetwork.evolve(myTrainingSet, {
generations: 10,
selection: methods.selection.TOURNAMENT // eg.
});
- Source:
- Default Value:
{ "name": "TOURNAMENT", "size": 5, "probability": 0.5 }