Lab 06: Auctions

This lab introduces auction theory through competitive bidding mechanisms.

Game Overview

Type: Auction-based games Players: 2+ players Rounds: Multiple rounds with bidding phases Stages: Multi-stage with bidding and allocation phases

Games

Auction Game

Actions: Bid amounts for items
State Space: Current bids, valuations, and auction history
Key Concept: Optimal bidding strategies and auction design

State Space

Observations

observation = {
    "current_round": 3,         # Current auction round
    "my_valuation": 50,         # My value for the item
    "current_bids": [30, 45],   # All current bids
    "auction_history": [...],   # Previous auction results
    "items_remaining": 5        # Items left to auction
}

Actions

action = 25  # Bid amount (integer)
action = 0   # No bid

Rewards

Auction outcome payoffs:

# Second-price auction example
if my_bid == max_bid:  # I win
    second_highest = sorted(bids)[-2]
    reward = my_valuation - second_highest
else:  # I don't win
    reward = 0

Game Structure

Stage Type

Multi-stage with bidding and allocation phases
Auction dynamics - competitive bidding environment
Valuation-based - different players have different values

Learning Opportunities

Optimal bidding - bid optimally given valuations
Auction design - understand different auction formats
Strategic behavior - avoid overbidding and underbidding

Testing

Local Testing

from core.engine import Engine
from core.game.AuctionGame import AuctionGame
from core.agents.lab06.random_agent import RandomAgent

my_agent = MyAgent("MyAgent")
opponent = RandomAgent("Random")

engine = Engine(AuctionGame(), [my_agent, opponent], rounds=100)
results = engine.run()

print(f"My score: {results[0]}")
print(f"Opponent score: {results[1]}")

Bidding Analysis

def analyze_bidding(self):
    if hasattr(self, 'bid_history'):
        avg_bid = sum(self.bid_history) / len(self.bid_history)
        print(f"Average bid: {avg_bid}")
        print(f"Bid range: {min(self.bid_history)} - {max(self.bid_history)}")

Next Steps

Implement an auction agent using the common patterns
Study auction theory to understand optimal strategies
Test different bidding approaches against various opponents
Compete against other students

Focus on understanding auction mechanisms and optimal bidding strategies!