gen-alg

The library of genetic algorithm

$3

``javascript
const { Population } = require('gen-alg');

const population = new Population({
schema: {
a: {
type: 'float',
min: 0,
max: 1,
digits: 2,
},
},
toFitness: entity => 1 - ((entity.a - 0.5) ** 2),
});

let stop = false;
for (let i = 0; i < 1000 && !stop; i++) {
stop = population.next(); // next generation
const { a } = population.best();
console.log(${i}. - ${a});
}


`

$3

`javascript

const schema = {
field1: {
type: 'int',
min: 0,
max: 100,
},
field2: {
type: 'float',
min: 0,
max: 1,
digits: 3,
},
field3: {
type: 'object',
fields: {
field1: {
type: 'int',
min: 0,
max: 50,
}
}
},
};

`

$3

`javascript
/**
* Gets fitness value
* @param {number} x genetic field
* @param {number} value target value
* @param {number} se standard error (se = Math.sqrt(max - min))
* @return {number} fitness value [0, 1]
*/
function forValue(x, value, se) {
if (x === value) return 1;
const dev = Math.abs(x - value) / se;
return Math.exp(-dev);
}
`
for example:
`javascript
const { Population, fitnessHelper } = require('gen-alg');

const population = new Population({
schema: {
a: {
type: 'int',
min: 0,
max: 100,
},
},
// size: 20,
toFitness: entity => fitnessHelper.forValue(entity.a, 50, 10),
});
let stop = false;
for (let i = 0; i < 1000 && !stop; i++) {
stop = population.next();
const { a } = population.best();
console.log(${i}. - ${a});
}

`

$3

`javascript
const { Population, fitnessHelper } = require('gen-alg');

const population = new Population({
schema: {
a: {
type: 'int',
min: 0,
max: 100,
},
b: {
type: 'float',
min: 0,
max: 1,
digits: 2,
},
},
size: 40, // you can performance manipulate
toFitness: ({ a, b }) => fitnessHelper.forValue(a, 50, 10) * fitnessHelper.forValue(b, 0.5, 0.1),
});
let stop = false;
for (let i = 0; i < 1000 && !stop; i++) {
stop = population.next();
const { a, b } = population.best();
console.log(${i}. - ${a}, ${b});
}

`

$3

The array type doesn't support yet. If you want use array you can write little hack,
like that:

`javascript
const { Population, fitnessHelper } = require('gen-alg');

// ------ Array hack ------
const length = 9;
const arraySchema = {};
for (let i = 0; i < length; i++) {
arraySchema[i] = {
type: 'int',
min: 0,
max: 100,
};
}

function toArray(entity) {
const arr = [];
for (let i = 0; i < length; i++) {
arr.push(entity[i]);
}
return arr;
}
// -----------------------

const population = new Population({
schema: arraySchema,
toFitness: (entity) => {
let fitness = 1;
const arr = toArray(entity);
for (let i = 0; i < arr.length; i++) {
const x = arr[i];
const targetValue = (i + 1) * 10;
fitness *= fitnessHelper.forValue(x, targetValue, 10);
}
return fitness;
},
});

let stop = false;
for (let i = 0; i < 1000 && !stop; i++) {
stop = population.next();
const arr = toArray(population.best());
console.log(${i}. - ${arr.toString()});
}

``