SirStone/SirRoboGarage
1
0
Fork 0
Code Issues 28 Pull Requests Actions Projects Releases Wiki Activity

feat(BNNBot): shaped reward based on miss distance — every shot teaches something

Browse Source 
Replace binary hit/miss reward with exponential decay based on miss distance:
- reward = 2.0 * exp(-missDistance / 36.0) - 1.0
- At 0px: +1.0 (perfect hit)
- At 36px: -0.26 (near miss, small penalty)
- At 100px: -0.87 (big miss, large penalty)

Maintains virtual hit/miss counters for display (threshold: 36px).

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
This commit is contained in:
Davide Cappellini
2026-09-18 00:10:28 +02:00
parent 63685b1729
commit f12bdf810e
1 changed files with 9 additions and 4 deletions
Download Patch File Download Diff File Expand all files Collapse all files
BNNBot_garage/src/BNNBot.nim
+9 -4
View File
@@ -96,12 +96,17 @@ method onScannedBot*(bot: BNNBot, e: ScannedBotEvent) =
if bulletDist >= b.fireDist or b.trace.age >= TRACE_MAX_AGE: if bulletDist >= b.fireDist or b.trace.age >= TRACE_MAX_AGE:
let bulletX = b.fireX + cos(degToRad(b.aimAngleDeg)) * bulletDist let bulletX = b.fireX + cos(degToRad(b.aimAngleDeg)) * bulletDist
let bulletY = b.fireY + sin(degToRad(b.aimAngleDeg)) * bulletDist let bulletY = b.fireY + sin(degToRad(b.aimAngleDeg)) * bulletDist
let enemyDist = hypot(bulletX - bot.lastEnemyX, bulletY - bot.lastEnemyY) let missDistance = hypot(bulletX - bot.lastEnemyX, bulletY - bot.lastEnemyY)
if enemyDist < 36.0:
bot.net.learn(b.trace, HIT_REWARD) # Shaped reward: +1.0 for perfect hit, decays toward -1.0 as miss distance grows
# Using exponential decay: reward = 2.0 * exp(-missDistance / 36.0) - 1.0
let reward = 2.0 * exp(-missDistance / 36.0) - 1.0
bot.net.learn(b.trace, reward)
# Track hit/miss for display: hit if within 36px, miss otherwise
if missDistance < 36.0:
inc bot.virtualHits inc bot.virtualHits
else: else:
bot.net.learn(b.trace, MISS_PENALTY)
inc bot.virtualMiss inc bot.virtualMiss
b.active = false b.active = false