Lab 07: Advanced Auctions

This lab introduces advanced auction mechanisms and bidding strategies.

Game Overview

Type: Multi-player auction games Players: 2+ players Rounds: Multiple rounds with complex bidding phases Stages: Multi-stage with bidding, allocation, and payment phases

Games

Advanced Auction Game

Actions: Complex bid structures with multiple parameters
State Space: Market segments, valuations, and auction history
Key Concept: Advanced bidding strategies and auction design

State Space

Observations

observation = {
    "current_round": 3,         # Current auction round
    "market_segments": [...],   # Available market segments
    "my_valuations": {...},     # My valuations for each segment
    "auction_history": [...],   # Previous auction results
    "remaining_budget": 1000,   # My remaining budget
    "opponent_bids": {...}      # Other players' bid information
}

Actions

action = {
    "bid_entries": [
        {
            "market_segment": "MALE_YOUNG_HIGH_INCOME",
            "bid": 2.5,
            "spending_limit": 100
        }
    ],
    "day_limit": 500
}

Rewards

Complex auction payoffs:

# Profit calculation
profit = (impressions_won / target_reach) * campaign_budget - total_spent
reward = profit

Game Structure

Stage Type

Multi-stage with complex bidding and allocation
Market segmentation - different segments have different values
Budget constraints - limited spending across segments

Learning Opportunities

Optimal bidding - balance bid amounts and spending limits
Market analysis - understand segment valuations
Budget management - allocate budget across segments efficiently

Testing

Local Testing

from core.engine import Engine
from core.game.AuctionGame import AuctionGame
from core.agents.lab07.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]}")

Budget Analysis

def analyze_budget_usage(self):
    if hasattr(self, 'spending_history'):
        total_spent = sum(self.spending_history)
        avg_spent = total_spent / len(self.spending_history)
        print(f"Average spending per round: {avg_spent}")
        print(f"Budget utilization: {total_spent / self.initial_budget:.2%}")

Next Steps

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

Focus on understanding advanced auction mechanisms and optimal bidding strategies!