Skip to content

Instantly share code, notes, and snippets.

@cauerego

cauerego/index.html

Created February 23, 2017 12:36
Show Gist options
  • Save cauerego/38d5159a191c96621700df713b16d37a to your computer and use it in GitHub Desktop.
Save cauerego/38d5159a191c96621700df713b16d37a to your computer and use it in GitHub Desktop.
Download ZIP
JS Bin google-hashcode2017-pizza // source https://jsbin.com/cadezov
Raw
index.html
<!DOCTYPE html>
<html>
<head>
<meta name="description" content="google-hashcode2017-pizza">
<meta charset="utf-8">
<meta name="viewport" content="width=device-width">
<title>JS Bin</title>
</head>
<body>
<script id="jsbin-javascript">
/*
for now this is mostly parts taken from mariox many files
it's also being adjusted to the pizza exercise
// */
//*
ActualInputs = [];
ActualOutputs = [];
Inputs = ActualInputs.length;
Outputs = ActualOutputs.length;
Population = 300; // species
DeltaDisjoint = 2.0;
DeltaWeights = 0.4;
DeltaThreshold = 1.0;
StaleSpecies = 15;
MutateConnectionsChance = 0.25;
PerturbChance = 0.90;
CrossoverChance = 0.75;
LinkMutationChance = 2.0;
NodeMutationChance = 0.50;
BiasMutationChance = 0.40;
StepSize = 0.1;
DisableMutationChance = 0.4;
EnableMutationChance = 0.2;
TimeoutConstant = 20;
Timeout = 6; // seconds
// ok, not quite a toolbox yet, not even a singleton, but we'll get there (if needed) (from toolbox.js)
var pool;
var rightmost;
var markedCells;
var timeout;
var timeElapsed;
var score;
var goal;
var solution;
if ( isEmpty(pool) ) {
initializePool();
}
// those are currently in the "global" scope, but only being used here (from main.js)
var fpsinterval = 0;
var mainLoopInterval = null;
var keepTime = null;
function pizza (str) {
var lines = str.split('\n');
var config = lines[0].split(' ');
lines.shift();
goal = config[0] * config[1];
var minIngredients = config[2]; // per slice
var maxSlice = config[3]; // cells size
// like the mario screen, this should probably morph into a different screen as the game progress.
// maybe just setting -1 to cells that are already taken would be enough
ActualInputs = [];
lines.forEach(function(item){
ActualInputs += item.replace(/M/g, 1).replace(/T/g, 0);
});
// still need to figure out what the outputs need to be, and probably adjust the rest of neat around it
// right now this is meaningless and just as a place holder
ActualOutputs = [
' ',
'0',
'1',
'3',
'4',
'5',
'6',
'7',
'8',
'9', // last command is being ignored, right now! :(
];
Inputs = goal+1;
Outputs = ActualOutputs.length;
startMainLoop();
}
function startMainLoop () {
mainLoopInterval = setInterval(asyncMainLoop, fpsinterval);
keepTime = setInterval(function(){ timeElapsed += 1; }, 1000);
}
function logScore () {
console.log(
'score: '+ score,
'goal: '+ goal,
'time elapsed: '+ timeElapsed,
'time limit: '+ Timeout,
'\n'+ solution
);
}
function clearMainLoop () {
logScore();
clearInterval(mainLoopInterval);
clearInterval(keepTime);
}
// still no good error handling, sadly
/*window.onerror = function(msg, url, line, col, error) {
// Note that col & error are new to the HTML 5 spec and may not be
// supported in every browser. It worked for me in Chrome.
var extra = !col ? '' : '\ncolumn: ' + col;
extra += !error ? '' : '\nerror: ' + error;
// You can view the information in an alert to see things working like this:
console.error("error :o " + msg + "\nurl: " + url + "\nline: " + line + extra);
clearMainLoop();
// TODO: Report this error via ajax so you can keep track
// of what pages have JS issues
var suppressErrorAlert = true;
// If you return true, then error alerts (like in older versions of
// Internet Explorer) will be suppressed.
return suppressErrorAlert;
};*/
function asyncMainLoop () { // infinite, async equivalent
if (score >= goal) {
console.log('goal reached! :)');
clearMainLoop();
}
if (timeElapsed > Timeout) {
console.log('time limit '+ Timeout +' reached! :(');
clearMainLoop();
}
// try {
aiMainLoop();
/* } catch (e) {
console.error('error :o '+ e);
clearMainLoop();
}*/
}
function aiMainLoop () {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
if ( (pool.currentFrame % 5) === 0 ) {
evaluateCurrent();
}
calculateNewInputs();
if (markedCells > rightmost) {
rightmost = markedCells;
timeout = TimeoutConstant;
}
timeout = timeout - 1;
var timeoutBonus = pool.currentFrame / 4;
if (timeout + timeoutBonus <= 0) {
var fitness = rightmost - pool.currentFrame / 2;
if (fitness === 0) {
fitness = -1;
}
genome.fitness = fitness;
if (fitness > pool.maxFitness) {
pool.maxFitness = fitness;
// $form.find('input#maxFitness').val(Math.floor(pool.maxFitness));
// writeFile( "autobackup.fitness." + fitness + "." + $form.find('input#saveLoadFile').val() );
// writeFile("autobackup.pool");
}
console.log(
"gen: " + pool.generation,
"species: " + pool.currentSpecies,
"genome: " + pool.currentGenome,
"fitness: " + fitness
);
logScore();
pool.currentSpecies = 0; // array bonds
pool.currentGenome = 0; // array bonds
while ( fitnessAlreadyMeasured() ) {
nextGenome();
}
initializeRun();
}
var measured = 0;
var total = 0;
Object.keys(pool.species).forEach( function(sKey) { // in pairs
var pairSpecies = pool.species[sKey];
Object.keys(pairSpecies.genomes).forEach( function (gKey) { // in pairs
var pairGenome = pairSpecies.genomes[gKey];
total++;
if (pairGenome.fitness !== 0) {
measured++;
}
});
});
/*
$aigui.find('#banner #gen').text( pool.generation + ' species ' + pool.currentSpecies + ' genome ' + pool.currentGenome + ' (' + Math.floor(measured/total*100) + '%)' );
$aigui.find('#banner #fitness').text( Math.floor(rightmost - (pool.currentFrame) / 2 - (timeout + timeoutBonus)*2/3) );
$aigui.find('#banner #maxFitness').text( Math.floor(pool.maxFitness) );
// */
pool.currentFrame++;
}
function sigmoid (x) {
return 2/(1+Math.exp(-4.9*x))-1;
}
function newPool () {
var pool = {};
pool.species = [];
pool.generation = 0;
pool.innovation = Outputs - 1; // array bonds
pool.currentSpecies = 0; // array bonds
pool.currentGenome = 0; // array bonds
pool.currentFrame = 0;
pool.maxFitness = 0;
pool.duration = 0;
pool.gameState = null;
pool.state = null;
return pool;
}
function newSpecies () {
var species = {};
species.topFitness = 0;
species.staleness = 0;
species.genomes = [];
species.averageFitness = 0;
return species;
}
function newGenome () {
var genome = {};
genome.genes = [];
genome.fitness = 0;
genome.adjustedFitness = 0;
genome.network = [];
genome.maxneuron = 0;
genome.globalRank = 0;
genome.mutationRates = {};
genome.mutationRates.connections = MutateConnectionsChance;
genome.mutationRates.link = LinkMutationChance;
genome.mutationRates.bias = BiasMutationChance;
genome.mutationRates.node = NodeMutationChance;
genome.mutationRates.enable = EnableMutationChance;
genome.mutationRates.disable = DisableMutationChance;
genome.mutationRates.step = StepSize;
return genome;
}
function copyGenome (genome) {
var genome2 = newGenome();
//for (var g=0; g<genome.genes.length; g++) {
// genome2.genes.push( copyGene(genome.genes[g]) ); // table.insert
genome.genes.forEach( function (gene) {
genome2.genes.push( copyGene(gene) ); // table.insert
});
genome2.maxneuron = genome.maxneuron;
genome2.mutationRates.connections = genome.mutationRates.connections;
genome2.mutationRates.link = genome.mutationRates.link;
genome2.mutationRates.bias = genome.mutationRates.bias;
genome2.mutationRates.node = genome.mutationRates.node;
genome2.mutationRates.enable = genome.mutationRates.enable;
genome2.mutationRates.disable = genome.mutationRates.disable;
return genome2;
}
function basicGenome () {
var genome = newGenome();
//var innovation = 0; // array bonds - probably useless
genome.maxneuron = Inputs - 1; // array bonds
mutate(genome);
return genome;
}
function newGene () {
var gene = {};
gene.into = 0;
gene.out = 0;
gene.weight = 0.0;
gene.enabled = true;
gene.innovation = 0;
return gene;
}
function copyGene (gene) {
var gene2 = newGene();
gene2.into = gene.into;
gene2.out = gene.out;
gene2.weight = gene.weight;
gene2.enabled = gene.enabled;
gene2.innovation = gene.innovation;
return gene2;
}
function newNeuron () {
var neuron = {};
neuron.incoming = [];
neuron.value = 0.0;
return neuron;
}
function generateNetwork (genome) {
var network = {};
network.inNeurons = [];
network.outNeurons = [];
for (var i=0; i<Inputs; i++) {
network.inNeurons[i] = newNeuron();
}
for (var o=0; o<Outputs; o++) {
network.outNeurons[o] = newNeuron();
}
genome.genes.sort(function (a, b) {
return (a.out - b.out);
})
function checkGene (gene, neurons) {
if (gene.enabled) {
if ( isEmpty(neurons[gene.out]) ) {
neurons[gene.out] = newNeuron();
}
var neuron = neurons[gene.out];
neuron.incoming.push(gene); // table.insert
if ( isEmpty(neurons[gene.into]) ) {
neurons[gene.into] = newNeuron();
}
}
}
genome.genes.forEach( function(gene) {
checkGene(gene, network.inNeurons);
checkGene(gene, network.outNeurons);
});
genome.network = network;
}
function evaluateNetwork (network, inputs) {
var outputs = {};
inputs.push(1); // table.insert
if (inputs.length != Inputs) {
console.error("Incorrect number of neural network inputs: "+ inputs.length +" (expected "+ Inputs +")");
return outputs;
}
for (var i=0; i<Inputs; i++) {
network.inNeurons[i].value = inputs[i];
}
var forEachNeuron = function (neuron) { // in pairs
var sum = 0;
for (var j = 0; j<neuron.incoming.length; j++) {
var incoming = neuron.incoming[j];
var other = network.inNeurons[incoming.into];
sum = sum + incoming.weight * other.value;
}
if (neuron.incoming.length > 0) {
neuron.value = sigmoid(sum);
}
}
network.inNeurons.forEach(forEachNeuron);
network.outNeurons.forEach(forEachNeuron);
for (var o=0; o<Outputs; o++) {
var outputName = "KEY_" + ActualOutputs[o];
if (network.outNeurons[o].value > 0) {
outputs[outputName] = true;
} else {
outputs[outputName] = false;
}
}
return outputs;
}
function crossover (g1, g2) {
// Make sure g1 is the higher fitness genome
if (g2.fitness > g1.fitness) {
tempg = g1;
g1 = g2;
g2 = tempg;
}
var child = newGenome();
var innovations2 = {};
for (var i=0; i<g2.genes.length; i++) {
var gene = g2.genes[i];
innovations2[gene.innovation] = gene;
}
for (var i=0; i<g1.genes.length; i++) {
var gene1 = g1.genes[i];
var gene2 = innovations2[gene1.innovation];
if ( !isEmpty(gene2) && mathRandom(2) == 1 && gene2.enabled) {
child.genes.push( copyGene(gene2) ); // table.insert
} else {
child.genes.push( copyGene(gene1) ); // table.insert
}
}
child.maxneuron = Math.max(g1.maxneuron,g2.maxneuron);
for (var mutation in g1.mutationRates) { // in pairs
var rate = g1.mutationRates[mutation];
child.mutationRates[mutation] = rate;
}
return child;
}
function randomNeuron (genes, nonInput) {
var neurons = [];
if ( !nonInput ) {
for (var i=0; i<Inputs; i++) {
neurons[i] = true;
}
}
for (var o=0; o<Outputs; o++) {
neurons[MaxNodes+o] = true;
}
for (var i=0; i<genes.length; i++) {
if ( !nonInput || genes[i].into >= Inputs) {
neurons[genes[i].into] = true;
}
if ( !nonInput || genes[i].out >= Inputs) {
neurons[genes[i].out] = true;
}
}
var count = 0;
for (var _ in neurons) { // in pairs
count = count + 1;
}
var n = mathRandom(1, count);
for (var k in neurons) { // in pairs
var v = neurons[k];
n = n-1;
if (n === 0) {
return k;
}
}
return 0;
}
function containsLink (genes, link) {
for (var i=0; i<genes.length; i++) {
var gene = genes[i];
if (gene.into == link.into && gene.out == link.out) {
return true;
}
}
}
function pointMutate (genome) {
var step = genome.mutationRates["step"];
for (var i=0; i<genome.genes.length; i++) {
var gene = genome.genes[i];
if (mathRandom() < PerturbChance) {
gene.weight = gene.weight + mathRandom() * step*2 - step;
} else {
gene.weight = mathRandom()*4-2;
}
}
}
function linkMutate (genome, forceBias) {
var neuron1 = randomNeuron(genome.genes, false);
var neuron2 = randomNeuron(genome.genes, true);
var newLink = newGene();
if (neuron1 < Inputs && neuron2 < Inputs) { // array bonds
// Both input nodes
return;
}
if (neuron2 < Inputs) { // array bonds
// Swap output and input
var temp = neuron1;
neuron1 = neuron2;
neuron2 = temp;
}
newLink.into = neuron1;
newLink.out = neuron2;
if (forceBias) {
newLink.into = Inputs - 1; // array bonds
}
if ( containsLink(genome.genes, newLink) ) {
return;
}
newLink.innovation = ++pool.innovation;
newLink.weight = mathRandom()*4-2;
genome.genes.push(newLink); // table.insert
}
function nodeMutate (genome) {
if (genome.genes.length === 0) {
return;
}
genome.maxneuron++;
var gene = genome.genes[mathRandom(1,genome.genes.length)-1];
if ( !gene || !gene.enabled ) {
return;
}
gene.enabled = false;
var gene1 = copyGene(gene);
gene1.out = genome.maxneuron;
gene1.weight = 1.0;
gene1.innovation = ++pool.innovation;
gene1.enabled = true;
genome.genes.push(gene1); // table.insert
var gene2 = copyGene(gene);
gene2.into = genome.maxneuron;
gene2.innovation = ++pool.innovation;
gene2.enabled = true;
genome.genes.push(gene2); // table.insert
}
function enableDisableMutate (genome, enable) {
var candidates = [];
for (var _ in genome.genes) { // in pairs
var gene = genome.genes[_];
if (gene.enabled == !enable) {
candidates.push(gene); // table.insert
}
}
if (candidates.length === 0) {
return;
}
var gene = candidates[mathRandom(1,candidates.length)-1];
gene.enabled = !gene.enabled;
}
function mutate (genome) {
for (var mutation in genome.mutationRates) { // in pairs
var rate = genome.mutationRates[mutation];
if (mathRandom(1,2) == 1) {
genome.mutationRates[mutation] = 0.95*rate;
} else {
genome.mutationRates[mutation] = 1.05263*rate;
}
}
if (mathRandom() < genome.mutationRates["connections"]) {
pointMutate(genome);
}
var p = genome.mutationRates["link"];
while (p > 0) {
if (mathRandom() < p) {
linkMutate(genome, false);
}
p--;
}
p = genome.mutationRates["bias"];
while (p > 0) {
if (mathRandom() < p) {
linkMutate(genome, true);
}
p--;
}
p = genome.mutationRates["node"];
while (p > 0) {
if (mathRandom() < p) {
nodeMutate(genome);
}
p--;
}
p = genome.mutationRates["enable"]
while (p > 0) {
if (mathRandom() < p) {
enableDisableMutate(genome, true);
}
p--;
}
p = genome.mutationRates["disable"]
while (p > 0) {
if (mathRandom() < p) {
enableDisableMutate(genome, false);
}
p--;
}
}
function disjoint (genes1, genes2) {
var i1 = [];
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
i1[gene.innovation] = true;
}
var i2 = [];
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
i2[gene.innovation] = true;
}
var disjointGenes = 0;
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
if (!i2[gene.innovation]) {
disjointGenes = disjointGenes+1;
}
}
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
if (!i1[gene.innovation]) {
disjointGenes = disjointGenes+1;
}
}
var n = Math.max(genes1.length-1, genes2.length-1);
return disjointGenes / n;
}
function weights (genes1, genes2) {
var i2 = [];
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
i2[gene.innovation] = gene;
}
var sum = 0;
var coincident = 0;
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
if ( !isEmpty(i2[gene.innovation]) ) {
var gene2 = i2[gene.innovation];
sum = sum + Math.abs(gene.weight - gene2.weight);
coincident++;
}
}
return sum / coincident;
}
function sameSpecies (genome1, genome2) {
var dd = DeltaDisjoint*disjoint(genome1.genes, genome2.genes);
var dw = DeltaWeights*weights(genome1.genes, genome2.genes);
return dd + dw < DeltaThreshold;
}
function rankGlobally () {
var global = [];
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
for (var g = 0; g <species.genomes.length; g ++) {
global.push(species.genomes[g]); // table.insert
}
}
global.sort(function (a, b) {
return (a.fitness - b.fitness); // from less to more fit
})
for (var g=0; g<global.length; g++) {
global[g].globalRank = g;
}
}
function calculateAverageFitness (species) {
var total = 0;
for (var g=0; g<species.genomes.length; g++) {
var genome = species.genomes[g];
total = total + genome.globalRank;
}
species.averageFitness = total / species.genomes.length;
}
function totalAverageFitness () {
var total = 0;
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
total = total + species.averageFitness;
}
return total;
}
function cullSpecies (cutToOne) {
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
species.genomes.sort(function (a, b) {
return (b.fitness - a.fitness);
})
var remaining = Math.ceil(species.genomes.length/2);
if (cutToOne) {
remaining = 1; // array bonds
}
while (species.genomes.length > remaining) {
species.genomes.pop();
}
}
}
function breedChild (species) {
var child = {};
if (mathRandom() < CrossoverChance) {
g1 = species.genomes[mathRandom(1, species.genomes.length)-1];
g2 = species.genomes[mathRandom(1, species.genomes.length)-1];
child = crossover(g1, g2);
} else {
g = species.genomes[mathRandom(1, species.genomes.length)-1];
child = copyGenome(g);
}
mutate(child);
return child;
}
function removeStaleSpecies () {
var survived = [];
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
species.genomes.sort(function (a, b) {
return (b.fitness - a.fitness);
})
if (species.genomes[0].fitness > species.topFitness) { // array bonds
species.topFitness = species.genomes[0].fitness; // array bonds
species.staleness = 0;
} else {
species.staleness++;
}
if (species.staleness < StaleSpecies || species.topFitness >= pool.maxFitness) {
survived.push(species); // table.insert
}
}
pool.species = survived;
}
function removeWeakSpecies () {
var survived = [];
var sum = totalAverageFitness();
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
var breed = Math.floor(species.averageFitness / sum * Population);
if (breed >= 1) {
survived.push(species); // table.insert
}
}
pool.species = survived;
}
function addToSpecies (child) {
var foundSpecies = false;
for (var s=0; s<pool.species.length; s++) {
var species = pool.species[s];
if ( !foundSpecies && sameSpecies(child, species.genomes[0]) ) { // array bonds
species.genomes.push(child); // table.insert
foundSpecies = true;
break; //for
}
}
if (!foundSpecies) {
var childSpecies = newSpecies();
childSpecies.genomes.push(child); // table.insert
pool.species.push(childSpecies); // table.insert
}
}
function newGeneration () {
cullSpecies(false); // Cull the bottom half of each species
rankGlobally();
removeStaleSpecies();
rankGlobally();
for (var s = 0; s<pool.species.length; s++) {
var species = pool.species[s];
calculateAverageFitness(species);
}
removeWeakSpecies();
var sum = totalAverageFitness();
var children = [];
for (var s = 0; s<pool.species.length; s++) {
var species = pool.species[s];
var breed = Math.floor(species.averageFitness / sum * Population) - 1;
for (var i=0; i<breed; i++) {
children.push( breedChild(species) ); // table.insert
}
}
cullSpecies(true); // Cull all but the top member of each species
while (children.length + pool.species.length <= Population) {
var species = pool.species[mathRandom(1, pool.species.length)-1];
children.push( breedChild(species) ); // table.insert
}
for (var c=0; c<children.length; c++) {
var child = children[c];
addToSpecies(child);
}
pool.generation++;
// writeFile("autobackup.gen." + pool.generation + "." + $form.find('input#saveLoadFile').val());
// writeFile("autobackup.pool");
}
function initializePool () {
pool = newPool();
for (var i=0; i<Population; i++) {
var basic = basicGenome();
addToSpecies(basic);
}
initializeRun();
}
function clearJoypad () {
/* controller = {};
for (var b = 0; b<ActualOutputs.length; b++) {
controller["KEY_" + ActualOutputs[b]] = false;
}*/
// joypadSet(controller);
}
function initializeRun () {
// review - something like savestate will be much needed
//loadState(Filename);
rightmost = 0;
pool.currentFrame = 0;
timeout = TimeoutConstant;
clearJoypad();
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
generateNetwork(genome);
evaluateCurrent();
}
function evaluateCurrent() {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
inputs = ActualInputs.slice();
var outputs = evaluateNetwork(genome.network, inputs);
/*
inputs = getInputs();
controller = evaluateNetwork(genome.network, inputs);
if (controller["KEY_LEFT"] && controller["KEY_RIGHT"]) {
controller["KEY_LEFT"] = false;
controller["KEY_RIGHT"] = false;
}
if (controller["KEY_UP"] && controller["KEY_DOWN"]) {
controller["KEY_UP"] = false;
controller["KEY_DOWN"] = false;
}
// */
}
function nextGenome () {
pool.currentGenome++;
if (pool.currentGenome >= pool.species[pool.currentSpecies].genomes.length) {
pool.currentGenome = 0; // array bonds
pool.currentSpecies++;
if (pool.currentSpecies >= pool.species.length) {
newGeneration();
pool.currentSpecies = 0; // array bonds
}
}
}
function fitnessAlreadyMeasured () {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
return genome.fitness !== 0;
}
function playTop () {
var maxFitness = 0;
var maxSpecies = 0;
var maxGenome = 0;
for (var s in pool.species) { // in pairs
var species = pool.species[s];
for (var g in species.genomes) { // in pairs
var genome = species.genomes[g];
if (genome.fitness > maxFitness) {
maxFitness = genome.fitness;
maxSpecies = s;
maxGenome = g;
}
}
}
pool.currentSpecies = maxSpecies;
pool.currentGenome = maxGenome;
pool.maxFitness = maxFitness;
$form.find('input#maxFitness').val(Math.floor(pool.maxFitness));
initializeRun();
pool.currentFrame++;
return;
}
// adapted functions to work like lua script (from lua.js)
function mathRandom (min, max) {
if ( isEmpty(min) ) {
return Math.random();
}
if ( isEmpty(max) ) {
max = min;
min = 1;
}
return Math.floor(Math.random() * (max - min)) + min;
}
function isEmpty (foo) {
return (foo == null); // should work for undefined as well
}
function getInputs () {
var inputs = ActualInputs.slice();
/* getPositions();
var sprites = getSprites();
var inputs = [];
for (var dy=-BoxRadius*16; dy<=BoxRadius*16; dy+=16) {
for (var dx=-BoxRadius*16; dx<=BoxRadius*16; dx+=16) {
inputs[inputs.length+0] = 0; // array bonds
tile = getTile(dx, dy);
if (tile == 1 && marioY+dy < 0x1B0) {
inputs[inputs.length-1] = 1; // array bonds
}
for (var i = 0; i<sprites.length; i++) { // array bonds
distx = Math.abs(sprites[i].x - (marioX+dx));
disty = Math.abs(sprites[i].y - (marioY+dy));
if (distx <= 8 && disty <= 8) {
inputs[inputs.length-1] = -1; // array bonds
}
}
}
}*/
return inputs;
}
function calculateNewInputs () {
// grabs the current output and mix up with the inputs to see how many markedCells are there
markedCells += 1; // dummy test
}
// */
var example = "3 5 1 6\nTTTTT\nTMMMT\nTTTTT\n";
var small = "6 7 1 5\nTMMMTTT\nMMMMTMM\nTTMTTMT\nTMMTMMM\nTTTTTTM\nTTTTTTM\n";
console.log( pizza(example) ); // 15
// console.log( pizza(small) ); // 42
</script>
<script id="jsbin-source-javascript" type="text/javascript">/*
for now this is mostly parts taken from mariox many files
it's also being adjusted to the pizza exercise
// */
//*
ActualInputs = [];
ActualOutputs = [];
Inputs = ActualInputs.length;
Outputs = ActualOutputs.length;
Population = 300; // species
DeltaDisjoint = 2.0;
DeltaWeights = 0.4;
DeltaThreshold = 1.0;
StaleSpecies = 15;
MutateConnectionsChance = 0.25;
PerturbChance = 0.90;
CrossoverChance = 0.75;
LinkMutationChance = 2.0;
NodeMutationChance = 0.50;
BiasMutationChance = 0.40;
StepSize = 0.1;
DisableMutationChance = 0.4;
EnableMutationChance = 0.2;
TimeoutConstant = 20;
Timeout = 6; // seconds
// ok, not quite a toolbox yet, not even a singleton, but we'll get there (if needed) (from toolbox.js)
var pool;
var rightmost;
var markedCells;
var timeout;
var timeElapsed;
var score;
var goal;
var solution;
if ( isEmpty(pool) ) {
initializePool();
}
// those are currently in the "global" scope, but only being used here (from main.js)
var fpsinterval = 0;
var mainLoopInterval = null;
var keepTime = null;
function pizza (str) {
var lines = str.split('\n');
var config = lines[0].split(' ');
lines.shift();
goal = config[0] * config[1];
var minIngredients = config[2]; // per slice
var maxSlice = config[3]; // cells size
// like the mario screen, this should probably morph into a different screen as the game progress.
// maybe just setting -1 to cells that are already taken would be enough
ActualInputs = [];
lines.forEach(function(item){
ActualInputs += item.replace(/M/g, 1).replace(/T/g, 0);
});
// still need to figure out what the outputs need to be, and probably adjust the rest of neat around it
// right now this is meaningless and just as a place holder
ActualOutputs = [
' ',
'0',
'1',
'3',
'4',
'5',
'6',
'7',
'8',
'9', // last command is being ignored, right now! :(
];
Inputs = goal+1;
Outputs = ActualOutputs.length;
startMainLoop();
}
function startMainLoop () {
mainLoopInterval = setInterval(asyncMainLoop, fpsinterval);
keepTime = setInterval(function(){ timeElapsed += 1; }, 1000);
}
function logScore () {
console.log(
'score: '+ score,
'goal: '+ goal,
'time elapsed: '+ timeElapsed,
'time limit: '+ Timeout,
'\n'+ solution
);
}
function clearMainLoop () {
logScore();
clearInterval(mainLoopInterval);
clearInterval(keepTime);
}
// still no good error handling, sadly
/*window.onerror = function(msg, url, line, col, error) {
// Note that col & error are new to the HTML 5 spec and may not be
// supported in every browser. It worked for me in Chrome.
var extra = !col ? '' : '\ncolumn: ' + col;
extra += !error ? '' : '\nerror: ' + error;
// You can view the information in an alert to see things working like this:
console.error("error :o " + msg + "\nurl: " + url + "\nline: " + line + extra);
clearMainLoop();
// TODO: Report this error via ajax so you can keep track
// of what pages have JS issues
var suppressErrorAlert = true;
// If you return true, then error alerts (like in older versions of
// Internet Explorer) will be suppressed.
return suppressErrorAlert;
};*/
function asyncMainLoop () { // infinite, async equivalent
if (score >= goal) {
console.log('goal reached! :)');
clearMainLoop();
}
if (timeElapsed > Timeout) {
console.log('time limit '+ Timeout +' reached! :(');
clearMainLoop();
}
// try {
aiMainLoop();
/* } catch (e) {
console.error('error :o '+ e);
clearMainLoop();
}*/
}
function aiMainLoop () {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
if ( (pool.currentFrame % 5) === 0 ) {
evaluateCurrent();
}
calculateNewInputs();
if (markedCells > rightmost) {
rightmost = markedCells;
timeout = TimeoutConstant;
}
timeout = timeout - 1;
var timeoutBonus = pool.currentFrame / 4;
if (timeout + timeoutBonus <= 0) {
var fitness = rightmost - pool.currentFrame / 2;
if (fitness === 0) {
fitness = -1;
}
genome.fitness = fitness;
if (fitness > pool.maxFitness) {
pool.maxFitness = fitness;
// $form.find('input#maxFitness').val(Math.floor(pool.maxFitness));
// writeFile( "autobackup.fitness." + fitness + "." + $form.find('input#saveLoadFile').val() );
// writeFile("autobackup.pool");
}
console.log(
"gen: " + pool.generation,
"species: " + pool.currentSpecies,
"genome: " + pool.currentGenome,
"fitness: " + fitness
);
logScore();
pool.currentSpecies = 0; // array bonds
pool.currentGenome = 0; // array bonds
while ( fitnessAlreadyMeasured() ) {
nextGenome();
}
initializeRun();
}
var measured = 0;
var total = 0;
Object.keys(pool.species).forEach( function(sKey) { // in pairs
var pairSpecies = pool.species[sKey];
Object.keys(pairSpecies.genomes).forEach( function (gKey) { // in pairs
var pairGenome = pairSpecies.genomes[gKey];
total++;
if (pairGenome.fitness !== 0) {
measured++;
}
});
});
/*
$aigui.find('#banner #gen').text( pool.generation + ' species ' + pool.currentSpecies + ' genome ' + pool.currentGenome + ' (' + Math.floor(measured/total*100) + '%)' );
$aigui.find('#banner #fitness').text( Math.floor(rightmost - (pool.currentFrame) / 2 - (timeout + timeoutBonus)*2/3) );
$aigui.find('#banner #maxFitness').text( Math.floor(pool.maxFitness) );
// */
pool.currentFrame++;
}
function sigmoid (x) {
return 2/(1+Math.exp(-4.9*x))-1;
}
function newPool () {
var pool = {};
pool.species = [];
pool.generation = 0;
pool.innovation = Outputs - 1; // array bonds
pool.currentSpecies = 0; // array bonds
pool.currentGenome = 0; // array bonds
pool.currentFrame = 0;
pool.maxFitness = 0;
pool.duration = 0;
pool.gameState = null;
pool.state = null;
return pool;
}
function newSpecies () {
var species = {};
species.topFitness = 0;
species.staleness = 0;
species.genomes = [];
species.averageFitness = 0;
return species;
}
function newGenome () {
var genome = {};
genome.genes = [];
genome.fitness = 0;
genome.adjustedFitness = 0;
genome.network = [];
genome.maxneuron = 0;
genome.globalRank = 0;
genome.mutationRates = {};
genome.mutationRates.connections = MutateConnectionsChance;
genome.mutationRates.link = LinkMutationChance;
genome.mutationRates.bias = BiasMutationChance;
genome.mutationRates.node = NodeMutationChance;
genome.mutationRates.enable = EnableMutationChance;
genome.mutationRates.disable = DisableMutationChance;
genome.mutationRates.step = StepSize;
return genome;
}
function copyGenome (genome) {
var genome2 = newGenome();
//for (var g=0; g<genome.genes.length; g++) {
// genome2.genes.push( copyGene(genome.genes[g]) ); // table.insert
genome.genes.forEach( function (gene) {
genome2.genes.push( copyGene(gene) ); // table.insert
});
genome2.maxneuron = genome.maxneuron;
genome2.mutationRates.connections = genome.mutationRates.connections;
genome2.mutationRates.link = genome.mutationRates.link;
genome2.mutationRates.bias = genome.mutationRates.bias;
genome2.mutationRates.node = genome.mutationRates.node;
genome2.mutationRates.enable = genome.mutationRates.enable;
genome2.mutationRates.disable = genome.mutationRates.disable;
return genome2;
}
function basicGenome () {
var genome = newGenome();
//var innovation = 0; // array bonds - probably useless
genome.maxneuron = Inputs - 1; // array bonds
mutate(genome);
return genome;
}
function newGene () {
var gene = {};
gene.into = 0;
gene.out = 0;
gene.weight = 0.0;
gene.enabled = true;
gene.innovation = 0;
return gene;
}
function copyGene (gene) {
var gene2 = newGene();
gene2.into = gene.into;
gene2.out = gene.out;
gene2.weight = gene.weight;
gene2.enabled = gene.enabled;
gene2.innovation = gene.innovation;
return gene2;
}
function newNeuron () {
var neuron = {};
neuron.incoming = [];
neuron.value = 0.0;
return neuron;
}
function generateNetwork (genome) {
var network = {};
network.inNeurons = [];
network.outNeurons = [];
for (var i=0; i<Inputs; i++) {
network.inNeurons[i] = newNeuron();
}
for (var o=0; o<Outputs; o++) {
network.outNeurons[o] = newNeuron();
}
genome.genes.sort(function (a, b) {
return (a.out - b.out);
})
function checkGene (gene, neurons) {
if (gene.enabled) {
if ( isEmpty(neurons[gene.out]) ) {
neurons[gene.out] = newNeuron();
}
var neuron = neurons[gene.out];
neuron.incoming.push(gene); // table.insert
if ( isEmpty(neurons[gene.into]) ) {
neurons[gene.into] = newNeuron();
}
}
}
genome.genes.forEach( function(gene) {
checkGene(gene, network.inNeurons);
checkGene(gene, network.outNeurons);
});
genome.network = network;
}
function evaluateNetwork (network, inputs) {
var outputs = {};
inputs.push(1); // table.insert
if (inputs.length != Inputs) {
console.error("Incorrect number of neural network inputs: "+ inputs.length +" (expected "+ Inputs +")");
return outputs;
}
for (var i=0; i<Inputs; i++) {
network.inNeurons[i].value = inputs[i];
}
var forEachNeuron = function (neuron) { // in pairs
var sum = 0;
for (var j = 0; j<neuron.incoming.length; j++) {
var incoming = neuron.incoming[j];
var other = network.inNeurons[incoming.into];
sum = sum + incoming.weight * other.value;
}
if (neuron.incoming.length > 0) {
neuron.value = sigmoid(sum);
}
}
network.inNeurons.forEach(forEachNeuron);
network.outNeurons.forEach(forEachNeuron);
for (var o=0; o<Outputs; o++) {
var outputName = "KEY_" + ActualOutputs[o];
if (network.outNeurons[o].value > 0) {
outputs[outputName] = true;
} else {
outputs[outputName] = false;
}
}
return outputs;
}
function crossover (g1, g2) {
// Make sure g1 is the higher fitness genome
if (g2.fitness > g1.fitness) {
tempg = g1;
g1 = g2;
g2 = tempg;
}
var child = newGenome();
var innovations2 = {};
for (var i=0; i<g2.genes.length; i++) {
var gene = g2.genes[i];
innovations2[gene.innovation] = gene;
}
for (var i=0; i<g1.genes.length; i++) {
var gene1 = g1.genes[i];
var gene2 = innovations2[gene1.innovation];
if ( !isEmpty(gene2) && mathRandom(2) == 1 && gene2.enabled) {
child.genes.push( copyGene(gene2) ); // table.insert
} else {
child.genes.push( copyGene(gene1) ); // table.insert
}
}
child.maxneuron = Math.max(g1.maxneuron,g2.maxneuron);
for (var mutation in g1.mutationRates) { // in pairs
var rate = g1.mutationRates[mutation];
child.mutationRates[mutation] = rate;
}
return child;
}
function randomNeuron (genes, nonInput) {
var neurons = [];
if ( !nonInput ) {
for (var i=0; i<Inputs; i++) {
neurons[i] = true;
}
}
for (var o=0; o<Outputs; o++) {
neurons[MaxNodes+o] = true;
}
for (var i=0; i<genes.length; i++) {
if ( !nonInput || genes[i].into >= Inputs) {
neurons[genes[i].into] = true;
}
if ( !nonInput || genes[i].out >= Inputs) {
neurons[genes[i].out] = true;
}
}
var count = 0;
for (var _ in neurons) { // in pairs
count = count + 1;
}
var n = mathRandom(1, count);
for (var k in neurons) { // in pairs
var v = neurons[k];
n = n-1;
if (n === 0) {
return k;
}
}
return 0;
}
function containsLink (genes, link) {
for (var i=0; i<genes.length; i++) {
var gene = genes[i];
if (gene.into == link.into && gene.out == link.out) {
return true;
}
}
}
function pointMutate (genome) {
var step = genome.mutationRates["step"];
for (var i=0; i<genome.genes.length; i++) {
var gene = genome.genes[i];
if (mathRandom() < PerturbChance) {
gene.weight = gene.weight + mathRandom() * step*2 - step;
} else {
gene.weight = mathRandom()*4-2;
}
}
}
function linkMutate (genome, forceBias) {
var neuron1 = randomNeuron(genome.genes, false);
var neuron2 = randomNeuron(genome.genes, true);
var newLink = newGene();
if (neuron1 < Inputs && neuron2 < Inputs) { // array bonds
// Both input nodes
return;
}
if (neuron2 < Inputs) { // array bonds
// Swap output and input
var temp = neuron1;
neuron1 = neuron2;
neuron2 = temp;
}
newLink.into = neuron1;
newLink.out = neuron2;
if (forceBias) {
newLink.into = Inputs - 1; // array bonds
}
if ( containsLink(genome.genes, newLink) ) {
return;
}
newLink.innovation = ++pool.innovation;
newLink.weight = mathRandom()*4-2;
genome.genes.push(newLink); // table.insert
}
function nodeMutate (genome) {
if (genome.genes.length === 0) {
return;
}
genome.maxneuron++;
var gene = genome.genes[mathRandom(1,genome.genes.length)-1];
if ( !gene || !gene.enabled ) {
return;
}
gene.enabled = false;
var gene1 = copyGene(gene);
gene1.out = genome.maxneuron;
gene1.weight = 1.0;
gene1.innovation = ++pool.innovation;
gene1.enabled = true;
genome.genes.push(gene1); // table.insert
var gene2 = copyGene(gene);
gene2.into = genome.maxneuron;
gene2.innovation = ++pool.innovation;
gene2.enabled = true;
genome.genes.push(gene2); // table.insert
}
function enableDisableMutate (genome, enable) {
var candidates = [];
for (var _ in genome.genes) { // in pairs
var gene = genome.genes[_];
if (gene.enabled == !enable) {
candidates.push(gene); // table.insert
}
}
if (candidates.length === 0) {
return;
}
var gene = candidates[mathRandom(1,candidates.length)-1];
gene.enabled = !gene.enabled;
}
function mutate (genome) {
for (var mutation in genome.mutationRates) { // in pairs
var rate = genome.mutationRates[mutation];
if (mathRandom(1,2) == 1) {
genome.mutationRates[mutation] = 0.95*rate;
} else {
genome.mutationRates[mutation] = 1.05263*rate;
}
}
if (mathRandom() < genome.mutationRates["connections"]) {
pointMutate(genome);
}
var p = genome.mutationRates["link"];
while (p > 0) {
if (mathRandom() < p) {
linkMutate(genome, false);
}
p--;
}
p = genome.mutationRates["bias"];
while (p > 0) {
if (mathRandom() < p) {
linkMutate(genome, true);
}
p--;
}
p = genome.mutationRates["node"];
while (p > 0) {
if (mathRandom() < p) {
nodeMutate(genome);
}
p--;
}
p = genome.mutationRates["enable"]
while (p > 0) {
if (mathRandom() < p) {
enableDisableMutate(genome, true);
}
p--;
}
p = genome.mutationRates["disable"]
while (p > 0) {
if (mathRandom() < p) {
enableDisableMutate(genome, false);
}
p--;
}
}
function disjoint (genes1, genes2) {
var i1 = [];
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
i1[gene.innovation] = true;
}
var i2 = [];
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
i2[gene.innovation] = true;
}
var disjointGenes = 0;
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
if (!i2[gene.innovation]) {
disjointGenes = disjointGenes+1;
}
}
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
if (!i1[gene.innovation]) {
disjointGenes = disjointGenes+1;
}
}
var n = Math.max(genes1.length-1, genes2.length-1);
return disjointGenes / n;
}
function weights (genes1, genes2) {
var i2 = [];
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
i2[gene.innovation] = gene;
}
var sum = 0;
var coincident = 0;
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
if ( !isEmpty(i2[gene.innovation]) ) {
var gene2 = i2[gene.innovation];
sum = sum + Math.abs(gene.weight - gene2.weight);
coincident++;
}
}
return sum / coincident;
}
function sameSpecies (genome1, genome2) {
var dd = DeltaDisjoint*disjoint(genome1.genes, genome2.genes);
var dw = DeltaWeights*weights(genome1.genes, genome2.genes);
return dd + dw < DeltaThreshold;
}
function rankGlobally () {
var global = [];
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
for (var g = 0; g <species.genomes.length; g ++) {
global.push(species.genomes[g]); // table.insert
}
}
global.sort(function (a, b) {
return (a.fitness - b.fitness); // from less to more fit
})
for (var g=0; g<global.length; g++) {
global[g].globalRank = g;
}
}
function calculateAverageFitness (species) {
var total = 0;
for (var g=0; g<species.genomes.length; g++) {
var genome = species.genomes[g];
total = total + genome.globalRank;
}
species.averageFitness = total / species.genomes.length;
}
function totalAverageFitness () {
var total = 0;
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
total = total + species.averageFitness;
}
return total;
}
function cullSpecies (cutToOne) {
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
species.genomes.sort(function (a, b) {
return (b.fitness - a.fitness);
})
var remaining = Math.ceil(species.genomes.length/2);
if (cutToOne) {
remaining = 1; // array bonds
}
while (species.genomes.length > remaining) {
species.genomes.pop();
}
}
}
function breedChild (species) {
var child = {};
if (mathRandom() < CrossoverChance) {
g1 = species.genomes[mathRandom(1, species.genomes.length)-1];
g2 = species.genomes[mathRandom(1, species.genomes.length)-1];
child = crossover(g1, g2);
} else {
g = species.genomes[mathRandom(1, species.genomes.length)-1];
child = copyGenome(g);
}
mutate(child);
return child;
}
function removeStaleSpecies () {
var survived = [];
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
species.genomes.sort(function (a, b) {
return (b.fitness - a.fitness);
})
if (species.genomes[0].fitness > species.topFitness) { // array bonds
species.topFitness = species.genomes[0].fitness; // array bonds
species.staleness = 0;
} else {
species.staleness++;
}
if (species.staleness < StaleSpecies || species.topFitness >= pool.maxFitness) {
survived.push(species); // table.insert
}
}
pool.species = survived;
}
function removeWeakSpecies () {
var survived = [];
var sum = totalAverageFitness();
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
var breed = Math.floor(species.averageFitness / sum * Population);
if (breed >= 1) {
survived.push(species); // table.insert
}
}
pool.species = survived;
}
function addToSpecies (child) {
var foundSpecies = false;
for (var s=0; s<pool.species.length; s++) {
var species = pool.species[s];
if ( !foundSpecies && sameSpecies(child, species.genomes[0]) ) { // array bonds
species.genomes.push(child); // table.insert
foundSpecies = true;
break; //for
}
}
if (!foundSpecies) {
var childSpecies = newSpecies();
childSpecies.genomes.push(child); // table.insert
pool.species.push(childSpecies); // table.insert
}
}
function newGeneration () {
cullSpecies(false); // Cull the bottom half of each species
rankGlobally();
removeStaleSpecies();
rankGlobally();
for (var s = 0; s<pool.species.length; s++) {
var species = pool.species[s];
calculateAverageFitness(species);
}
removeWeakSpecies();
var sum = totalAverageFitness();
var children = [];
for (var s = 0; s<pool.species.length; s++) {
var species = pool.species[s];
var breed = Math.floor(species.averageFitness / sum * Population) - 1;
for (var i=0; i<breed; i++) {
children.push( breedChild(species) ); // table.insert
}
}
cullSpecies(true); // Cull all but the top member of each species
while (children.length + pool.species.length <= Population) {
var species = pool.species[mathRandom(1, pool.species.length)-1];
children.push( breedChild(species) ); // table.insert
}
for (var c=0; c<children.length; c++) {
var child = children[c];
addToSpecies(child);
}
pool.generation++;
// writeFile("autobackup.gen." + pool.generation + "." + $form.find('input#saveLoadFile').val());
// writeFile("autobackup.pool");
}
function initializePool () {
pool = newPool();
for (var i=0; i<Population; i++) {
var basic = basicGenome();
addToSpecies(basic);
}
initializeRun();
}
function clearJoypad () {
/* controller = {};
for (var b = 0; b<ActualOutputs.length; b++) {
controller["KEY_" + ActualOutputs[b]] = false;
}*/
// joypadSet(controller);
}
function initializeRun () {
// review - something like savestate will be much needed
//loadState(Filename);
rightmost = 0;
pool.currentFrame = 0;
timeout = TimeoutConstant;
clearJoypad();
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
generateNetwork(genome);
evaluateCurrent();
}
function evaluateCurrent() {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
inputs = ActualInputs.slice();
var outputs = evaluateNetwork(genome.network, inputs);
/*
inputs = getInputs();
controller = evaluateNetwork(genome.network, inputs);
if (controller["KEY_LEFT"] && controller["KEY_RIGHT"]) {
controller["KEY_LEFT"] = false;
controller["KEY_RIGHT"] = false;
}
if (controller["KEY_UP"] && controller["KEY_DOWN"]) {
controller["KEY_UP"] = false;
controller["KEY_DOWN"] = false;
}
// */
}
function nextGenome () {
pool.currentGenome++;
if (pool.currentGenome >= pool.species[pool.currentSpecies].genomes.length) {
pool.currentGenome = 0; // array bonds
pool.currentSpecies++;
if (pool.currentSpecies >= pool.species.length) {
newGeneration();
pool.currentSpecies = 0; // array bonds
}
}
}
function fitnessAlreadyMeasured () {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
return genome.fitness !== 0;
}
function playTop () {
var maxFitness = 0;
var maxSpecies = 0;
var maxGenome = 0;
for (var s in pool.species) { // in pairs
var species = pool.species[s];
for (var g in species.genomes) { // in pairs
var genome = species.genomes[g];
if (genome.fitness > maxFitness) {
maxFitness = genome.fitness;
maxSpecies = s;
maxGenome = g;
}
}
}
pool.currentSpecies = maxSpecies;
pool.currentGenome = maxGenome;
pool.maxFitness = maxFitness;
$form.find('input#maxFitness').val(Math.floor(pool.maxFitness));
initializeRun();
pool.currentFrame++;
return;
}
// adapted functions to work like lua script (from lua.js)
function mathRandom (min, max) {
if ( isEmpty(min) ) {
return Math.random();
}
if ( isEmpty(max) ) {
max = min;
min = 1;
}
return Math.floor(Math.random() * (max - min)) + min;
}
function isEmpty (foo) {
return (foo == null); // should work for undefined as well
}
function getInputs () {
var inputs = ActualInputs.slice();
/* getPositions();
var sprites = getSprites();
var inputs = [];
for (var dy=-BoxRadius*16; dy<=BoxRadius*16; dy+=16) {
for (var dx=-BoxRadius*16; dx<=BoxRadius*16; dx+=16) {
inputs[inputs.length+0] = 0; // array bonds
tile = getTile(dx, dy);
if (tile == 1 && marioY+dy < 0x1B0) {
inputs[inputs.length-1] = 1; // array bonds
}
for (var i = 0; i<sprites.length; i++) { // array bonds
distx = Math.abs(sprites[i].x - (marioX+dx));
disty = Math.abs(sprites[i].y - (marioY+dy));
if (distx <= 8 && disty <= 8) {
inputs[inputs.length-1] = -1; // array bonds
}
}
}
}*/
return inputs;
}
function calculateNewInputs () {
// grabs the current output and mix up with the inputs to see how many markedCells are there
markedCells += 1; // dummy test
}
// */
var example = "3 5 1 6\nTTTTT\nTMMMT\nTTTTT\n";
var small = "6 7 1 5\nTMMMTTT\nMMMMTMM\nTTMTTMT\nTMMTMMM\nTTTTTTM\nTTTTTTM\n";
console.log( pizza(example) ); // 15
// console.log( pizza(small) ); // 42
</script></body>
</html>
Raw
jsbin.cadezov.js
/*
for now this is mostly parts taken from mariox many files
it's also being adjusted to the pizza exercise
// */
//*
ActualInputs = [];
ActualOutputs = [];
Inputs = ActualInputs.length;
Outputs = ActualOutputs.length;
Population = 300; // species
DeltaDisjoint = 2.0;
DeltaWeights = 0.4;
DeltaThreshold = 1.0;
StaleSpecies = 15;
MutateConnectionsChance = 0.25;
PerturbChance = 0.90;
CrossoverChance = 0.75;
LinkMutationChance = 2.0;
NodeMutationChance = 0.50;
BiasMutationChance = 0.40;
StepSize = 0.1;
DisableMutationChance = 0.4;
EnableMutationChance = 0.2;
TimeoutConstant = 20;
Timeout = 6; // seconds
// ok, not quite a toolbox yet, not even a singleton, but we'll get there (if needed) (from toolbox.js)
var pool;
var rightmost;
var markedCells;
var timeout;
var timeElapsed;
var score;
var goal;
var solution;
if ( isEmpty(pool) ) {
initializePool();
}
// those are currently in the "global" scope, but only being used here (from main.js)
var fpsinterval = 0;
var mainLoopInterval = null;
var keepTime = null;
function pizza (str) {
var lines = str.split('\n');
var config = lines[0].split(' ');
lines.shift();
goal = config[0] * config[1];
var minIngredients = config[2]; // per slice
var maxSlice = config[3]; // cells size
// like the mario screen, this should probably morph into a different screen as the game progress.
// maybe just setting -1 to cells that are already taken would be enough
ActualInputs = [];
lines.forEach(function(item){
ActualInputs += item.replace(/M/g, 1).replace(/T/g, 0);
});
// still need to figure out what the outputs need to be, and probably adjust the rest of neat around it
// right now this is meaningless and just as a place holder
ActualOutputs = [
' ',
'0',
'1',
'3',
'4',
'5',
'6',
'7',
'8',
'9', // last command is being ignored, right now! :(
];
Inputs = goal+1;
Outputs = ActualOutputs.length;
startMainLoop();
}
function startMainLoop () {
mainLoopInterval = setInterval(asyncMainLoop, fpsinterval);
keepTime = setInterval(function(){ timeElapsed += 1; }, 1000);
}
function logScore () {
console.log(
'score: '+ score,
'goal: '+ goal,
'time elapsed: '+ timeElapsed,
'time limit: '+ Timeout,
'\n'+ solution
);
}
function clearMainLoop () {
logScore();
clearInterval(mainLoopInterval);
clearInterval(keepTime);
}
// still no good error handling, sadly
/*window.onerror = function(msg, url, line, col, error) {
// Note that col & error are new to the HTML 5 spec and may not be
// supported in every browser. It worked for me in Chrome.
var extra = !col ? '' : '\ncolumn: ' + col;
extra += !error ? '' : '\nerror: ' + error;
// You can view the information in an alert to see things working like this:
console.error("error :o " + msg + "\nurl: " + url + "\nline: " + line + extra);
clearMainLoop();
// TODO: Report this error via ajax so you can keep track
// of what pages have JS issues
var suppressErrorAlert = true;
// If you return true, then error alerts (like in older versions of
// Internet Explorer) will be suppressed.
return suppressErrorAlert;
};*/
function asyncMainLoop () { // infinite, async equivalent
if (score >= goal) {
console.log('goal reached! :)');
clearMainLoop();
}
if (timeElapsed > Timeout) {
console.log('time limit '+ Timeout +' reached! :(');
clearMainLoop();
}
// try {
aiMainLoop();
/* } catch (e) {
console.error('error :o '+ e);
clearMainLoop();
}*/
}
function aiMainLoop () {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
if ( (pool.currentFrame % 5) === 0 ) {
evaluateCurrent();
}
calculateNewInputs();
if (markedCells > rightmost) {
rightmost = markedCells;
timeout = TimeoutConstant;
}
timeout = timeout - 1;
var timeoutBonus = pool.currentFrame / 4;
if (timeout + timeoutBonus <= 0) {
var fitness = rightmost - pool.currentFrame / 2;
if (fitness === 0) {
fitness = -1;
}
genome.fitness = fitness;
if (fitness > pool.maxFitness) {
pool.maxFitness = fitness;
// $form.find('input#maxFitness').val(Math.floor(pool.maxFitness));
// writeFile( "autobackup.fitness." + fitness + "." + $form.find('input#saveLoadFile').val() );
// writeFile("autobackup.pool");
}
console.log(
"gen: " + pool.generation,
"species: " + pool.currentSpecies,
"genome: " + pool.currentGenome,
"fitness: " + fitness
);
logScore();
pool.currentSpecies = 0; // array bonds
pool.currentGenome = 0; // array bonds
while ( fitnessAlreadyMeasured() ) {
nextGenome();
}
initializeRun();
}
var measured = 0;
var total = 0;
Object.keys(pool.species).forEach( function(sKey) { // in pairs
var pairSpecies = pool.species[sKey];
Object.keys(pairSpecies.genomes).forEach( function (gKey) { // in pairs
var pairGenome = pairSpecies.genomes[gKey];
total++;
if (pairGenome.fitness !== 0) {
measured++;
}
});
});
/*
$aigui.find('#banner #gen').text( pool.generation + ' species ' + pool.currentSpecies + ' genome ' + pool.currentGenome + ' (' + Math.floor(measured/total*100) + '%)' );
$aigui.find('#banner #fitness').text( Math.floor(rightmost - (pool.currentFrame) / 2 - (timeout + timeoutBonus)*2/3) );
$aigui.find('#banner #maxFitness').text( Math.floor(pool.maxFitness) );
// */
pool.currentFrame++;
}
function sigmoid (x) {
return 2/(1+Math.exp(-4.9*x))-1;
}
function newPool () {
var pool = {};
pool.species = [];
pool.generation = 0;
pool.innovation = Outputs - 1; // array bonds
pool.currentSpecies = 0; // array bonds
pool.currentGenome = 0; // array bonds
pool.currentFrame = 0;
pool.maxFitness = 0;
pool.duration = 0;
pool.gameState = null;
pool.state = null;
return pool;
}
function newSpecies () {
var species = {};
species.topFitness = 0;
species.staleness = 0;
species.genomes = [];
species.averageFitness = 0;
return species;
}
function newGenome () {
var genome = {};
genome.genes = [];
genome.fitness = 0;
genome.adjustedFitness = 0;
genome.network = [];
genome.maxneuron = 0;
genome.globalRank = 0;
genome.mutationRates = {};
genome.mutationRates.connections = MutateConnectionsChance;
genome.mutationRates.link = LinkMutationChance;
genome.mutationRates.bias = BiasMutationChance;
genome.mutationRates.node = NodeMutationChance;
genome.mutationRates.enable = EnableMutationChance;
genome.mutationRates.disable = DisableMutationChance;
genome.mutationRates.step = StepSize;
return genome;
}
function copyGenome (genome) {
var genome2 = newGenome();
//for (var g=0; g<genome.genes.length; g++) {
// genome2.genes.push( copyGene(genome.genes[g]) ); // table.insert
genome.genes.forEach( function (gene) {
genome2.genes.push( copyGene(gene) ); // table.insert
});
genome2.maxneuron = genome.maxneuron;
genome2.mutationRates.connections = genome.mutationRates.connections;
genome2.mutationRates.link = genome.mutationRates.link;
genome2.mutationRates.bias = genome.mutationRates.bias;
genome2.mutationRates.node = genome.mutationRates.node;
genome2.mutationRates.enable = genome.mutationRates.enable;
genome2.mutationRates.disable = genome.mutationRates.disable;
return genome2;
}
function basicGenome () {
var genome = newGenome();
//var innovation = 0; // array bonds - probably useless
genome.maxneuron = Inputs - 1; // array bonds
mutate(genome);
return genome;
}
function newGene () {
var gene = {};
gene.into = 0;
gene.out = 0;
gene.weight = 0.0;
gene.enabled = true;
gene.innovation = 0;
return gene;
}
function copyGene (gene) {
var gene2 = newGene();
gene2.into = gene.into;
gene2.out = gene.out;
gene2.weight = gene.weight;
gene2.enabled = gene.enabled;
gene2.innovation = gene.innovation;
return gene2;
}
function newNeuron () {
var neuron = {};
neuron.incoming = [];
neuron.value = 0.0;
return neuron;
}
function generateNetwork (genome) {
var network = {};
network.inNeurons = [];
network.outNeurons = [];
for (var i=0; i<Inputs; i++) {
network.inNeurons[i] = newNeuron();
}
for (var o=0; o<Outputs; o++) {
network.outNeurons[o] = newNeuron();
}
genome.genes.sort(function (a, b) {
return (a.out - b.out);
})
function checkGene (gene, neurons) {
if (gene.enabled) {
if ( isEmpty(neurons[gene.out]) ) {
neurons[gene.out] = newNeuron();
}
var neuron = neurons[gene.out];
neuron.incoming.push(gene); // table.insert
if ( isEmpty(neurons[gene.into]) ) {
neurons[gene.into] = newNeuron();
}
}
}
genome.genes.forEach( function(gene) {
checkGene(gene, network.inNeurons);
checkGene(gene, network.outNeurons);
});
genome.network = network;
}
function evaluateNetwork (network, inputs) {
var outputs = {};
inputs.push(1); // table.insert
if (inputs.length != Inputs) {
console.error("Incorrect number of neural network inputs: "+ inputs.length +" (expected "+ Inputs +")");
return outputs;
}
for (var i=0; i<Inputs; i++) {
network.inNeurons[i].value = inputs[i];
}
var forEachNeuron = function (neuron) { // in pairs
var sum = 0;
for (var j = 0; j<neuron.incoming.length; j++) {
var incoming = neuron.incoming[j];
var other = network.inNeurons[incoming.into];
sum = sum + incoming.weight * other.value;
}
if (neuron.incoming.length > 0) {
neuron.value = sigmoid(sum);
}
}
network.inNeurons.forEach(forEachNeuron);
network.outNeurons.forEach(forEachNeuron);
for (var o=0; o<Outputs; o++) {
var outputName = "KEY_" + ActualOutputs[o];
if (network.outNeurons[o].value > 0) {
outputs[outputName] = true;
} else {
outputs[outputName] = false;
}
}
return outputs;
}
function crossover (g1, g2) {
// Make sure g1 is the higher fitness genome
if (g2.fitness > g1.fitness) {
tempg = g1;
g1 = g2;
g2 = tempg;
}
var child = newGenome();
var innovations2 = {};
for (var i=0; i<g2.genes.length; i++) {
var gene = g2.genes[i];
innovations2[gene.innovation] = gene;
}
for (var i=0; i<g1.genes.length; i++) {
var gene1 = g1.genes[i];
var gene2 = innovations2[gene1.innovation];
if ( !isEmpty(gene2) && mathRandom(2) == 1 && gene2.enabled) {
child.genes.push( copyGene(gene2) ); // table.insert
} else {
child.genes.push( copyGene(gene1) ); // table.insert
}
}
child.maxneuron = Math.max(g1.maxneuron,g2.maxneuron);
for (var mutation in g1.mutationRates) { // in pairs
var rate = g1.mutationRates[mutation];
child.mutationRates[mutation] = rate;
}
return child;
}
function randomNeuron (genes, nonInput) {
var neurons = [];
if ( !nonInput ) {
for (var i=0; i<Inputs; i++) {
neurons[i] = true;
}
}
for (var o=0; o<Outputs; o++) {
neurons[MaxNodes+o] = true;
}
for (var i=0; i<genes.length; i++) {
if ( !nonInput || genes[i].into >= Inputs) {
neurons[genes[i].into] = true;
}
if ( !nonInput || genes[i].out >= Inputs) {
neurons[genes[i].out] = true;
}
}
var count = 0;
for (var _ in neurons) { // in pairs
count = count + 1;
}
var n = mathRandom(1, count);
for (var k in neurons) { // in pairs
var v = neurons[k];
n = n-1;
if (n === 0) {
return k;
}
}
return 0;
}
function containsLink (genes, link) {
for (var i=0; i<genes.length; i++) {
var gene = genes[i];
if (gene.into == link.into && gene.out == link.out) {
return true;
}
}
}
function pointMutate (genome) {
var step = genome.mutationRates["step"];
for (var i=0; i<genome.genes.length; i++) {
var gene = genome.genes[i];
if (mathRandom() < PerturbChance) {
gene.weight = gene.weight + mathRandom() * step*2 - step;
} else {
gene.weight = mathRandom()*4-2;
}
}
}
function linkMutate (genome, forceBias) {
var neuron1 = randomNeuron(genome.genes, false);
var neuron2 = randomNeuron(genome.genes, true);
var newLink = newGene();
if (neuron1 < Inputs && neuron2 < Inputs) { // array bonds
// Both input nodes
return;
}
if (neuron2 < Inputs) { // array bonds
// Swap output and input
var temp = neuron1;
neuron1 = neuron2;
neuron2 = temp;
}
newLink.into = neuron1;
newLink.out = neuron2;
if (forceBias) {
newLink.into = Inputs - 1; // array bonds
}
if ( containsLink(genome.genes, newLink) ) {
return;
}
newLink.innovation = ++pool.innovation;
newLink.weight = mathRandom()*4-2;
genome.genes.push(newLink); // table.insert
}
function nodeMutate (genome) {
if (genome.genes.length === 0) {
return;
}
genome.maxneuron++;
var gene = genome.genes[mathRandom(1,genome.genes.length)-1];
if ( !gene || !gene.enabled ) {
return;
}
gene.enabled = false;
var gene1 = copyGene(gene);
gene1.out = genome.maxneuron;
gene1.weight = 1.0;
gene1.innovation = ++pool.innovation;
gene1.enabled = true;
genome.genes.push(gene1); // table.insert
var gene2 = copyGene(gene);
gene2.into = genome.maxneuron;
gene2.innovation = ++pool.innovation;
gene2.enabled = true;
genome.genes.push(gene2); // table.insert
}
function enableDisableMutate (genome, enable) {
var candidates = [];
for (var _ in genome.genes) { // in pairs
var gene = genome.genes[_];
if (gene.enabled == !enable) {
candidates.push(gene); // table.insert
}
}
if (candidates.length === 0) {
return;
}
var gene = candidates[mathRandom(1,candidates.length)-1];
gene.enabled = !gene.enabled;
}
function mutate (genome) {
for (var mutation in genome.mutationRates) { // in pairs
var rate = genome.mutationRates[mutation];
if (mathRandom(1,2) == 1) {
genome.mutationRates[mutation] = 0.95*rate;
} else {
genome.mutationRates[mutation] = 1.05263*rate;
}
}
if (mathRandom() < genome.mutationRates["connections"]) {
pointMutate(genome);
}
var p = genome.mutationRates["link"];
while (p > 0) {
if (mathRandom() < p) {
linkMutate(genome, false);
}
p--;
}
p = genome.mutationRates["bias"];
while (p > 0) {
if (mathRandom() < p) {
linkMutate(genome, true);
}
p--;
}
p = genome.mutationRates["node"];
while (p > 0) {
if (mathRandom() < p) {
nodeMutate(genome);
}
p--;
}
p = genome.mutationRates["enable"]
while (p > 0) {
if (mathRandom() < p) {
enableDisableMutate(genome, true);
}
p--;
}
p = genome.mutationRates["disable"]
while (p > 0) {
if (mathRandom() < p) {
enableDisableMutate(genome, false);
}
p--;
}
}
function disjoint (genes1, genes2) {
var i1 = [];
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
i1[gene.innovation] = true;
}
var i2 = [];
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
i2[gene.innovation] = true;
}
var disjointGenes = 0;
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
if (!i2[gene.innovation]) {
disjointGenes = disjointGenes+1;
}
}
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
if (!i1[gene.innovation]) {
disjointGenes = disjointGenes+1;
}
}
var n = Math.max(genes1.length-1, genes2.length-1);
return disjointGenes / n;
}
function weights (genes1, genes2) {
var i2 = [];
for (var i = 0; i <genes2.length; i ++) {
var gene = genes2[i];
i2[gene.innovation] = gene;
}
var sum = 0;
var coincident = 0;
for (var i = 0; i <genes1.length; i ++) {
var gene = genes1[i];
if ( !isEmpty(i2[gene.innovation]) ) {
var gene2 = i2[gene.innovation];
sum = sum + Math.abs(gene.weight - gene2.weight);
coincident++;
}
}
return sum / coincident;
}
function sameSpecies (genome1, genome2) {
var dd = DeltaDisjoint*disjoint(genome1.genes, genome2.genes);
var dw = DeltaWeights*weights(genome1.genes, genome2.genes);
return dd + dw < DeltaThreshold;
}
function rankGlobally () {
var global = [];
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
for (var g = 0; g <species.genomes.length; g ++) {
global.push(species.genomes[g]); // table.insert
}
}
global.sort(function (a, b) {
return (a.fitness - b.fitness); // from less to more fit
})
for (var g=0; g<global.length; g++) {
global[g].globalRank = g;
}
}
function calculateAverageFitness (species) {
var total = 0;
for (var g=0; g<species.genomes.length; g++) {
var genome = species.genomes[g];
total = total + genome.globalRank;
}
species.averageFitness = total / species.genomes.length;
}
function totalAverageFitness () {
var total = 0;
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
total = total + species.averageFitness;
}
return total;
}
function cullSpecies (cutToOne) {
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
species.genomes.sort(function (a, b) {
return (b.fitness - a.fitness);
})
var remaining = Math.ceil(species.genomes.length/2);
if (cutToOne) {
remaining = 1; // array bonds
}
while (species.genomes.length > remaining) {
species.genomes.pop();
}
}
}
function breedChild (species) {
var child = {};
if (mathRandom() < CrossoverChance) {
g1 = species.genomes[mathRandom(1, species.genomes.length)-1];
g2 = species.genomes[mathRandom(1, species.genomes.length)-1];
child = crossover(g1, g2);
} else {
g = species.genomes[mathRandom(1, species.genomes.length)-1];
child = copyGenome(g);
}
mutate(child);
return child;
}
function removeStaleSpecies () {
var survived = [];
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
species.genomes.sort(function (a, b) {
return (b.fitness - a.fitness);
})
if (species.genomes[0].fitness > species.topFitness) { // array bonds
species.topFitness = species.genomes[0].fitness; // array bonds
species.staleness = 0;
} else {
species.staleness++;
}
if (species.staleness < StaleSpecies || species.topFitness >= pool.maxFitness) {
survived.push(species); // table.insert
}
}
pool.species = survived;
}
function removeWeakSpecies () {
var survived = [];
var sum = totalAverageFitness();
for (var s = 0; s <pool.species.length; s ++) {
var species = pool.species[s];
var breed = Math.floor(species.averageFitness / sum * Population);
if (breed >= 1) {
survived.push(species); // table.insert
}
}
pool.species = survived;
}
function addToSpecies (child) {
var foundSpecies = false;
for (var s=0; s<pool.species.length; s++) {
var species = pool.species[s];
if ( !foundSpecies && sameSpecies(child, species.genomes[0]) ) { // array bonds
species.genomes.push(child); // table.insert
foundSpecies = true;
break; //for
}
}
if (!foundSpecies) {
var childSpecies = newSpecies();
childSpecies.genomes.push(child); // table.insert
pool.species.push(childSpecies); // table.insert
}
}
function newGeneration () {
cullSpecies(false); // Cull the bottom half of each species
rankGlobally();
removeStaleSpecies();
rankGlobally();
for (var s = 0; s<pool.species.length; s++) {
var species = pool.species[s];
calculateAverageFitness(species);
}
removeWeakSpecies();
var sum = totalAverageFitness();
var children = [];
for (var s = 0; s<pool.species.length; s++) {
var species = pool.species[s];
var breed = Math.floor(species.averageFitness / sum * Population) - 1;
for (var i=0; i<breed; i++) {
children.push( breedChild(species) ); // table.insert
}
}
cullSpecies(true); // Cull all but the top member of each species
while (children.length + pool.species.length <= Population) {
var species = pool.species[mathRandom(1, pool.species.length)-1];
children.push( breedChild(species) ); // table.insert
}
for (var c=0; c<children.length; c++) {
var child = children[c];
addToSpecies(child);
}
pool.generation++;
// writeFile("autobackup.gen." + pool.generation + "." + $form.find('input#saveLoadFile').val());
// writeFile("autobackup.pool");
}
function initializePool () {
pool = newPool();
for (var i=0; i<Population; i++) {
var basic = basicGenome();
addToSpecies(basic);
}
initializeRun();
}
function clearJoypad () {
/* controller = {};
for (var b = 0; b<ActualOutputs.length; b++) {
controller["KEY_" + ActualOutputs[b]] = false;
}*/
// joypadSet(controller);
}
function initializeRun () {
// review - something like savestate will be much needed
//loadState(Filename);
rightmost = 0;
pool.currentFrame = 0;
timeout = TimeoutConstant;
clearJoypad();
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
generateNetwork(genome);
evaluateCurrent();
}
function evaluateCurrent() {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
inputs = ActualInputs.slice();
var outputs = evaluateNetwork(genome.network, inputs);
/*
inputs = getInputs();
controller = evaluateNetwork(genome.network, inputs);
if (controller["KEY_LEFT"] && controller["KEY_RIGHT"]) {
controller["KEY_LEFT"] = false;
controller["KEY_RIGHT"] = false;
}
if (controller["KEY_UP"] && controller["KEY_DOWN"]) {
controller["KEY_UP"] = false;
controller["KEY_DOWN"] = false;
}
// */
}
function nextGenome () {
pool.currentGenome++;
if (pool.currentGenome >= pool.species[pool.currentSpecies].genomes.length) {
pool.currentGenome = 0; // array bonds
pool.currentSpecies++;
if (pool.currentSpecies >= pool.species.length) {
newGeneration();
pool.currentSpecies = 0; // array bonds
}
}
}
function fitnessAlreadyMeasured () {
var species = pool.species[pool.currentSpecies];
var genome = species.genomes[pool.currentGenome];
return genome.fitness !== 0;
}
function playTop () {
var maxFitness = 0;
var maxSpecies = 0;
var maxGenome = 0;
for (var s in pool.species) { // in pairs
var species = pool.species[s];
for (var g in species.genomes) { // in pairs
var genome = species.genomes[g];
if (genome.fitness > maxFitness) {
maxFitness = genome.fitness;
maxSpecies = s;
maxGenome = g;
}
}
}
pool.currentSpecies = maxSpecies;
pool.currentGenome = maxGenome;
pool.maxFitness = maxFitness;
$form.find('input#maxFitness').val(Math.floor(pool.maxFitness));
initializeRun();
pool.currentFrame++;
return;
}
// adapted functions to work like lua script (from lua.js)
function mathRandom (min, max) {
if ( isEmpty(min) ) {
return Math.random();
}
if ( isEmpty(max) ) {
max = min;
min = 1;
}
return Math.floor(Math.random() * (max - min)) + min;
}
function isEmpty (foo) {
return (foo == null); // should work for undefined as well
}
function getInputs () {
var inputs = ActualInputs.slice();
/* getPositions();
var sprites = getSprites();
var inputs = [];
for (var dy=-BoxRadius*16; dy<=BoxRadius*16; dy+=16) {
for (var dx=-BoxRadius*16; dx<=BoxRadius*16; dx+=16) {
inputs[inputs.length+0] = 0; // array bonds
tile = getTile(dx, dy);
if (tile == 1 && marioY+dy < 0x1B0) {
inputs[inputs.length-1] = 1; // array bonds
}
for (var i = 0; i<sprites.length; i++) { // array bonds
distx = Math.abs(sprites[i].x - (marioX+dx));
disty = Math.abs(sprites[i].y - (marioY+dy));
if (distx <= 8 && disty <= 8) {
inputs[inputs.length-1] = -1; // array bonds
}
}
}
}*/
return inputs;
}
function calculateNewInputs () {
// grabs the current output and mix up with the inputs to see how many markedCells are there
markedCells += 1; // dummy test
}
// */
var example = "3 5 1 6\nTTTTT\nTMMMT\nTTTTT\n";
var small = "6 7 1 5\nTMMMTTT\nMMMMTMM\nTTMTTMT\nTMMTMMM\nTTTTTTM\nTTTTTTM\n";
console.log( pizza(example) ); // 15
// console.log( pizza(small) ); // 42
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment