File size: 16,303 Bytes
b18c436 5779997 |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 |
<!DOCTYPE html>
<html lang="en">
<head>
<meta charset="UTF-8" />
<meta name="viewport" content="width=device-width, initial-scale=1.0" />
<title>Simple Q-Learning Grid World Simulation</title>
<style>
body {
font-family: Arial, sans-serif;
max-width: 800px;
margin: 0 auto;
padding: 20px;
}
.grid {
display: grid;
grid-template-columns: repeat(4, 80px);
grid-template-rows: repeat(4, 80px);
gap: 2px;
margin: 20px 0;
}
.cell {
width: 80px;
height: 80px;
border: 1px solid #ccc;
display: flex;
align-items: center;
justify-content: center;
position: relative;
}
.agent {
width: 30px;
height: 30px;
background-color: blue;
border-radius: 50%;
position: absolute;
}
.goal {
background-color: green;
color: white;
}
.obstacle {
background-color: gray;
}
.controls {
margin: 20px 0;
}
button {
padding: 8px 16px;
margin-right: 10px;
cursor: pointer;
}
.info {
margin: 20px 0;
padding: 10px;
background-color: #f0f0f0;
border-radius: 5px;
}
.parameters {
display: grid;
grid-template-columns: auto 1fr auto;
gap: 10px;
align-items: center;
margin-bottom: 10px;
}
table {
border-collapse: collapse;
margin-top: 20px;
width: 100%;
}
th,
td {
border: 1px solid #ddd;
padding: 8px;
text-align: center;
}
.chart {
width: 100%;
height: 200px;
margin-top: 20px;
}
.signature {
text-align: center; /* Changed from 'right' to 'center' */
font-style: italic;
margin-top: 30px;
}
</style>
</head>
<body>
<h1>Simple Q-Learning Grid World Simulation - Designed by Pejman</h1>
<div class="info">
<p>
This simulation demonstrates Q-learning - a reinforcement learning
algorithm where an agent learns to navigate a grid world to reach a goal
while avoiding obstacles.
</p>
</div>
<div class="parameters">
<label for="alpha">Learning Rate (Ξ±):</label>
<input type="range" id="alpha" min="0.1" max="1" step="0.1" value="0.5" />
<span id="alpha-value">0.5</span>
<label for="gamma">Discount Factor (Ξ³):</label>
<input type="range" id="gamma" min="0.1" max="1" step="0.1" value="0.9" />
<span id="gamma-value">0.9</span>
<label for="epsilon">Exploration Rate (Ξ΅):</label>
<input type="range" id="epsilon" min="0" max="1" step="0.1" value="0.3" />
<span id="epsilon-value">0.3</span>
</div>
<div class="controls">
<button id="step-btn">Step</button>
<button id="train-btn">Train Episode</button>
<button id="auto-btn">Auto Train</button>
<button id="stop-btn" disabled>Stop</button>
<button id="reset-btn">Reset</button>
</div>
<div class="info" id="status">Episode: 1 | Step: 0 | Total Reward: 0</div>
<div class="grid" id="grid"></div>
<h2>Q-Table</h2>
<div id="q-table"></div>
<h2>Learning Progress</h2>
<canvas id="chart" class="chart"></canvas>
<div class="signature">
*Β© 2025 Pejman Ebrahimi - Basic Q-Learning Simulation*
</div>
<script src="https://cdn.jsdelivr.net/npm/chart.js"></script>
<script>
// Grid setup
const grid = document.getElementById("grid");
const gridSize = 4;
let agentPos = { x: 0, y: 0 };
const goalPos = { x: 3, y: 3 };
const obstacles = [
{ x: 1, y: 1 },
{ x: 2, y: 1 },
{ x: 1, y: 2 },
];
// Learning parameters
let alpha = 0.5;
let gamma = 0.9;
let epsilon = 0.3;
let qTable = {};
// Training variables
let episode = 1;
let step = 0;
let totalReward = 0;
let rewards = [];
let running = false;
// Actions
const actions = ["up", "right", "down", "left"];
// Initialize grid
function createGrid() {
grid.innerHTML = "";
for (let y = 0; y < gridSize; y++) {
for (let x = 0; x < gridSize; x++) {
const cell = document.createElement("div");
cell.className = "cell";
cell.id = `cell-${x}-${y}`;
if (x === goalPos.x && y === goalPos.y) {
cell.classList.add("goal");
cell.textContent = "GOAL";
} else if (obstacles.some((o) => o.x === x && o.y === y)) {
cell.classList.add("obstacle");
}
grid.appendChild(cell);
}
}
updateAgentPosition();
}
// Update agent position
function updateAgentPosition() {
const agent = document.querySelector(".agent");
if (agent) agent.remove();
const cell = document.getElementById(
`cell-${agentPos.x}-${agentPos.y}`
);
const agentElement = document.createElement("div");
agentElement.className = "agent";
cell.appendChild(agentElement);
}
// Initialize Q-Table
function initQTable() {
qTable = {};
for (let y = 0; y < gridSize; y++) {
for (let x = 0; x < gridSize; x++) {
if (obstacles.some((o) => o.x === x && o.y === y)) continue;
qTable[`${x},${y}`] = {
up: 0,
right: 0,
down: 0,
left: 0,
};
}
}
updateQTableDisplay();
}
// Update Q-Table display
function updateQTableDisplay() {
const tableContainer = document.getElementById("q-table");
tableContainer.innerHTML = "";
const table = document.createElement("table");
// Create header row
const thead = document.createElement("thead");
const headerRow = document.createElement("tr");
headerRow.appendChild(document.createElement("th"));
for (let x = 0; x < gridSize; x++) {
const th = document.createElement("th");
th.textContent = x;
headerRow.appendChild(th);
}
thead.appendChild(headerRow);
table.appendChild(thead);
// Create table body
const tbody = document.createElement("tbody");
for (let y = 0; y < gridSize; y++) {
const row = document.createElement("tr");
const th = document.createElement("th");
th.textContent = y;
row.appendChild(th);
for (let x = 0; x < gridSize; x++) {
const cell = document.createElement("td");
if (obstacles.some((o) => o.x === x && o.y === y)) {
cell.textContent = "X";
cell.style.backgroundColor = "lightgray";
} else if (x === goalPos.x && y === goalPos.y) {
cell.textContent = "GOAL";
cell.style.backgroundColor = "lightgreen";
} else {
const state = `${x},${y}`;
const stateQ = qTable[state];
// Find best action
let bestAction = actions[0];
let bestValue = stateQ[bestAction];
for (const action of actions) {
if (stateQ[action] > bestValue) {
bestValue = stateQ[action];
bestAction = action;
}
}
let actionSymbol = "";
switch (bestAction) {
case "up":
actionSymbol = "β";
break;
case "right":
actionSymbol = "β";
break;
case "down":
actionSymbol = "β";
break;
case "left":
actionSymbol = "β";
break;
}
cell.textContent = `${actionSymbol} (${bestValue.toFixed(1)})`;
// Color based on value
const normalizedValue = Math.max(
0,
Math.min(1, (bestValue + 5) / 10)
);
cell.style.backgroundColor = `rgba(0, 128, 0, ${
normalizedValue * 0.5
})`;
}
row.appendChild(cell);
}
tbody.appendChild(row);
}
table.appendChild(tbody);
tableContainer.appendChild(table);
}
// Choose action using epsilon-greedy policy
function chooseAction() {
const state = `${agentPos.x},${agentPos.y}`;
const validActions = getValidActions();
// Exploration
if (Math.random() < epsilon) {
return validActions[Math.floor(Math.random() * validActions.length)];
}
// Exploitation
const stateQ = qTable[state];
let bestAction = validActions[0];
let bestValue = stateQ[bestAction];
for (const action of validActions) {
if (stateQ[action] > bestValue) {
bestValue = stateQ[action];
bestAction = action;
}
}
return bestAction;
}
// Get valid actions for current state
function getValidActions() {
const validActions = [];
// Check up
if (agentPos.y > 0 && !isObstacle(agentPos.x, agentPos.y - 1)) {
validActions.push("up");
}
// Check right
if (
agentPos.x < gridSize - 1 &&
!isObstacle(agentPos.x + 1, agentPos.y)
) {
validActions.push("right");
}
// Check down
if (
agentPos.y < gridSize - 1 &&
!isObstacle(agentPos.x, agentPos.y + 1)
) {
validActions.push("down");
}
// Check left
if (agentPos.x > 0 && !isObstacle(agentPos.x - 1, agentPos.y)) {
validActions.push("left");
}
return validActions;
}
// Check if position is an obstacle
function isObstacle(x, y) {
return obstacles.some((o) => o.x === x && o.y === y);
}
// Take action and get reward
function takeAction(action) {
const oldPos = { ...agentPos };
// Update position based on action
switch (action) {
case "up":
agentPos.y = Math.max(0, agentPos.y - 1);
break;
case "right":
agentPos.x = Math.min(gridSize - 1, agentPos.x + 1);
break;
case "down":
agentPos.y = Math.min(gridSize - 1, agentPos.y + 1);
break;
case "left":
agentPos.x = Math.max(0, agentPos.x - 1);
break;
}
// Check if position is valid
if (isObstacle(agentPos.x, agentPos.y)) {
agentPos = oldPos;
return -10; // Hitting obstacle penalty
}
// Calculate reward
if (agentPos.x === goalPos.x && agentPos.y === goalPos.y) {
return 10; // Goal reward
}
return -1; // Step penalty
}
// Update Q-value for state-action pair
function updateQValue(state, action, reward, nextState) {
const currQ = qTable[state][action];
// Find max Q-value for next state
const nextStateQ = qTable[nextState];
const maxNextQ = Math.max(...Object.values(nextStateQ));
// Q-learning formula
const newQ = currQ + alpha * (reward + gamma * maxNextQ - currQ);
qTable[state][action] = newQ;
}
// Perform one training step
function performStep() {
const state = `${agentPos.x},${agentPos.y}`;
const action = chooseAction();
const reward = takeAction(action);
updateAgentPosition();
const nextState = `${agentPos.x},${agentPos.y}`;
updateQValue(state, action, reward, nextState);
step++;
totalReward += reward;
document.getElementById(
"status"
).textContent = `Episode: ${episode} | Step: ${step} | Total Reward: ${totalReward}`;
updateQTableDisplay();
// Check if episode is done
if (agentPos.x === goalPos.x && agentPos.y === goalPos.y) {
rewards.push(totalReward);
// Update chart
chart.data.labels.push(episode);
chart.data.datasets[0].data.push(totalReward);
chart.update();
// Start new episode
episode++;
resetAgentPosition();
return true; // Episode completed
}
return false; // Episode not completed
}
// Train a complete episode
function trainEpisode() {
let episodeDone = false;
while (!episodeDone) {
episodeDone = performStep();
}
}
// Auto-train function
function autoTrain() {
if (!running) return;
const episodeDone = performStep();
if (episodeDone) {
setTimeout(autoTrain, 200);
} else {
requestAnimationFrame(autoTrain);
}
}
// Reset agent position
function resetAgentPosition() {
agentPos = { x: 0, y: 0 };
updateAgentPosition();
step = 0;
totalReward = 0;
document.getElementById(
"status"
).textContent = `Episode: ${episode} | Step: ${step} | Total Reward: ${totalReward}`;
}
// Reset environment
function resetEnvironment() {
agentPos = { x: 0, y: 0 };
updateAgentPosition();
initQTable();
episode = 1;
step = 0;
totalReward = 0;
rewards = [];
document.getElementById(
"status"
).textContent = `Episode: ${episode} | Step: ${step} | Total Reward: ${totalReward}`;
// Reset chart
chart.data.labels = [];
chart.data.datasets[0].data = [];
chart.update();
}
// Initialize chart
const ctx = document.getElementById("chart").getContext("2d");
const chart = new Chart(ctx, {
type: "line",
data: {
labels: [],
datasets: [
{
label: "Total Reward",
data: [],
borderColor: "blue",
backgroundColor: "rgba(0, 0, 255, 0.1)",
tension: 0.1,
fill: true,
},
],
},
options: {
responsive: true,
scales: {
y: {
beginAtZero: false,
},
},
},
});
// Event listeners
document
.getElementById("step-btn")
.addEventListener("click", performStep);
document
.getElementById("train-btn")
.addEventListener("click", trainEpisode);
document
.getElementById("auto-btn")
.addEventListener("click", function () {
running = true;
this.disabled = true;
document.getElementById("stop-btn").disabled = false;
autoTrain();
});
document
.getElementById("stop-btn")
.addEventListener("click", function () {
running = false;
this.disabled = true;
document.getElementById("auto-btn").disabled = false;
});
document
.getElementById("reset-btn")
.addEventListener("click", resetEnvironment);
document.getElementById("alpha").addEventListener("input", function () {
alpha = parseFloat(this.value);
document.getElementById("alpha-value").textContent = alpha.toFixed(1);
});
document.getElementById("gamma").addEventListener("input", function () {
gamma = parseFloat(this.value);
document.getElementById("gamma-value").textContent = gamma.toFixed(1);
});
document.getElementById("epsilon").addEventListener("input", function () {
epsilon = parseFloat(this.value);
document.getElementById("epsilon-value").textContent =
epsilon.toFixed(1);
});
// Initialize environment
createGrid();
initQTable();
</script>
</body>
</html>
|