@kometbomb/genetic-algorithm
Version:
Simple genetic algorithm helper class. Supports asynchronous fitness evaluation.
246 lines (245 loc) • 13.5 kB
JavaScript
"use strict";
var __assign = (this && this.__assign) || function () {
__assign = Object.assign || function(t) {
for (var s, i = 1, n = arguments.length; i < n; i++) {
s = arguments[i];
for (var p in s) if (Object.prototype.hasOwnProperty.call(s, p))
t[p] = s[p];
}
return t;
};
return __assign.apply(this, arguments);
};
var __awaiter = (this && this.__awaiter) || function (thisArg, _arguments, P, generator) {
function adopt(value) { return value instanceof P ? value : new P(function (resolve) { resolve(value); }); }
return new (P || (P = Promise))(function (resolve, reject) {
function fulfilled(value) { try { step(generator.next(value)); } catch (e) { reject(e); } }
function rejected(value) { try { step(generator["throw"](value)); } catch (e) { reject(e); } }
function step(result) { result.done ? resolve(result.value) : adopt(result.value).then(fulfilled, rejected); }
step((generator = generator.apply(thisArg, _arguments || [])).next());
});
};
var __generator = (this && this.__generator) || function (thisArg, body) {
var _ = { label: 0, sent: function() { if (t[0] & 1) throw t[1]; return t[1]; }, trys: [], ops: [] }, f, y, t, g;
return g = { next: verb(0), "throw": verb(1), "return": verb(2) }, typeof Symbol === "function" && (g[Symbol.iterator] = function() { return this; }), g;
function verb(n) { return function (v) { return step([n, v]); }; }
function step(op) {
if (f) throw new TypeError("Generator is already executing.");
while (_) try {
if (f = 1, y && (t = op[0] & 2 ? y["return"] : op[0] ? y["throw"] || ((t = y["return"]) && t.call(y), 0) : y.next) && !(t = t.call(y, op[1])).done) return t;
if (y = 0, t) op = [op[0] & 2, t.value];
switch (op[0]) {
case 0: case 1: t = op; break;
case 4: _.label++; return { value: op[1], done: false };
case 5: _.label++; y = op[1]; op = [0]; continue;
case 7: op = _.ops.pop(); _.trys.pop(); continue;
default:
if (!(t = _.trys, t = t.length > 0 && t[t.length - 1]) && (op[0] === 6 || op[0] === 2)) { _ = 0; continue; }
if (op[0] === 3 && (!t || (op[1] > t[0] && op[1] < t[3]))) { _.label = op[1]; break; }
if (op[0] === 6 && _.label < t[1]) { _.label = t[1]; t = op; break; }
if (t && _.label < t[2]) { _.label = t[2]; _.ops.push(op); break; }
if (t[2]) _.ops.pop();
_.trys.pop(); continue;
}
op = body.call(thisArg, _);
} catch (e) { op = [6, e]; y = 0; } finally { f = t = 0; }
if (op[0] & 5) throw op[1]; return { value: op[0] ? op[1] : void 0, done: true };
}
};
exports.__esModule = true;
exports.GeneticAlgorithm = void 0;
var GeneticAlgorithm = /** @class */ (function () {
/**
* Construct a new GeneticAlgorithm. The type parameter is the type of the genotype passed to the callback functions.
*
* @param config Initial configuration. Can be changed runtime with .evolve()
* @param initialPopulation The initial population. If config.populationSize is larger than this value, the rest will be filled with versions of the initial population.
*/
function GeneticAlgorithm(config, initialPopulation) {
var _this = this;
this.setConfig = function (config) {
if (config.populationSize < 2) {
throw new Error("populationSize has to be greater than one.");
}
if (config.elitistRatio !== undefined && !(config.elitistRatio >= 0 && config.elitistRatio <= 1)) {
throw new Error("elititstRatio has to be between 0.0 and 1.0.");
}
_this.config = __assign({}, config);
return config;
};
this.getRankedPopulation = function (recalculateFitness) {
if (recalculateFitness === void 0) { recalculateFitness = false; }
return __awaiter(_this, void 0, void 0, function () {
var hasNoFitness, hasFitness, fitness_1, rankedPopulation;
return __generator(this, function (_a) {
switch (_a.label) {
case 0:
hasNoFitness = this.population.filter(function (ranked) { return typeof ranked.fitness !== "number" || recalculateFitness; });
hasFitness = this.population.filter(function (ranked) { return typeof ranked.fitness === "number" && !recalculateFitness; });
if (!(hasNoFitness.length > 0)) return [3 /*break*/, 2];
return [4 /*yield*/, this.config.fitnessFunction(hasNoFitness.map(function (ranked) { return ranked.genotype; }))];
case 1:
fitness_1 = _a.sent();
if (fitness_1.length !== hasNoFitness.length) {
throw new Error("fitnessFunction should return as many fitness values as there are input genotypes.");
}
rankedPopulation = hasNoFitness.map(function (ranked, index) { return ({ genotype: ranked.genotype, fitness: fitness_1[index] }); }).concat(hasFitness);
this.population = rankedPopulation;
return [2 /*return*/, rankedPopulation];
case 2: return [2 /*return*/, hasFitness];
}
});
});
};
this.crossover = function (phenotype, mate) {
return _this.config.crossoverFunction(phenotype, mate);
};
this.compete = function () { return __awaiter(_this, void 0, void 0, function () {
var crossoverProbability, rankedPopulation, total, accumulatedFitness, elitistRatio, nextGeneration, getRandomParent, a, b;
return __generator(this, function (_a) {
switch (_a.label) {
case 0:
crossoverProbability = this.config.crossoverProbability !== undefined ? this.config.crossoverProbability : 0.5;
return [4 /*yield*/, this.getRankedPopulation(!!this.config.recalculateFitnessBeforeEachGeneration)];
case 1:
rankedPopulation = (_a.sent())
.map(function (item) { return (__assign(__assign({}, item), { accumulatedFitness: 0 })); });
rankedPopulation.sort(function (a, b) { return b.fitness - a.fitness; });
total = rankedPopulation.reduce(function (prev, curr) { return prev + curr.fitness; }, 0) || 1;
accumulatedFitness = 0;
rankedPopulation = rankedPopulation.map(function (genotype) {
accumulatedFitness += genotype.fitness / total;
return __assign(__assign({}, genotype), { accumulatedFitness: accumulatedFitness });
});
elitistRatio = this.config.elitistRatio !== undefined ? this.config.elitistRatio : 0.25;
nextGeneration = rankedPopulation.slice(0, this.config.populationSize * elitistRatio);
getRandomParent = function () {
var r = Math.random();
var genotype = rankedPopulation.find(function (genotype) { return genotype.accumulatedFitness >= r; });
if (!genotype) {
return rankedPopulation[Math.floor(Math.random() * rankedPopulation.length)];
}
return genotype;
};
while (nextGeneration.length < this.config.populationSize) {
a = getRandomParent();
if (this.config.crossoverFunction && Math.random() < crossoverProbability) {
b = getRandomParent();
nextGeneration.push({ genotype: this.crossover(a.genotype, b.genotype), fitness: null });
}
else {
nextGeneration.push({ genotype: this.mutate(a.genotype), fitness: null });
}
}
// Cull back to populationSize
this.population = nextGeneration.slice(0, this.config.populationSize);
return [2 /*return*/];
}
});
}); };
this.mutate = function (genotype) {
return _this.config.mutationFunction(genotype);
};
this.populate = function () {
var size = _this.population.length;
while (_this.population.length < _this.config.populationSize) {
_this.population.push({
genotype: _this.mutate(_this.population[Math.floor(Math.random() * size)].genotype),
fitness: null
});
}
};
/**
* Run for one generation.
*
* @param config Optional config to replace the config given to the constructor
*/
this.evolve = function (config) { return __awaiter(_this, void 0, void 0, function () {
return __generator(this, function (_a) {
switch (_a.label) {
case 0:
if (config) {
this.setConfig(config);
}
this.populate();
return [4 /*yield*/, this.compete()];
case 1:
_a.sent();
return [2 /*return*/];
}
});
}); };
/**
* @returns The best ranked genotype with fitness value.
*/
this.bestRanked = function () { return __awaiter(_this, void 0, void 0, function () {
var ranked;
return __generator(this, function (_a) {
switch (_a.label) {
case 0: return [4 /*yield*/, this.getRankedPopulation()];
case 1:
ranked = (_a.sent()).filter(function (ranked) { return typeof ranked.fitness === "number"; });
if (ranked.length === 0) {
throw new Error("Could not find genotypes with a calculated fitness value - did you run .evolve() yet?");
}
return [2 /*return*/, ranked.sort(function (a, b) { return b.fitness - a.fitness; })[0]];
}
});
}); };
/**
* Only valid after running evolve().
*
* @returns The best ranked genotype.
*/
this.best = function () { return __awaiter(_this, void 0, void 0, function () {
return __generator(this, function (_a) {
switch (_a.label) {
case 0: return [4 /*yield*/, this.bestRanked()];
case 1: return [2 /*return*/, (_a.sent()).genotype];
}
});
}); };
/**
* Only valid after running evolve().
*
* @returns The best fitness value.
*/
this.bestScore = function () { return __awaiter(_this, void 0, void 0, function () {
return __generator(this, function (_a) {
switch (_a.label) {
case 0: return [4 /*yield*/, this.bestRanked()];
case 1: return [2 /*return*/, (_a.sent()).fitness];
}
});
}); };
/**
* @returns The full genotype population.
*/
this.getPopulation = function () {
return _this.population.map(function (ranked) { return ranked.genotype; });
};
/**
* Only valid after running evolve().
*
* @returns The mean fitness value.
*/
this.meanFitness = function () { return __awaiter(_this, void 0, void 0, function () {
var withFitness;
return __generator(this, function (_a) {
switch (_a.label) {
case 0: return [4 /*yield*/, this.getRankedPopulation()];
case 1:
withFitness = (_a.sent()).filter(function (ranked) { return ranked.fitness !== null; });
return [2 /*return*/, withFitness.reduce(function (acc, ranked) { return acc + ranked.fitness; }, 0) / withFitness.length];
}
});
}); };
if (initialPopulation.length < 1) {
throw new Error("Initial population has to be given.");
}
this.population = initialPopulation.map(function (genotype) { return ({ genotype: genotype, fitness: null }); });
this.config = this.setConfig(config);
}
return GeneticAlgorithm;
}());
exports.GeneticAlgorithm = GeneticAlgorithm;