Blackjack Strategy Evolution using Evolutionary Algorithms

The system uses three evolutionary approaches to discover Blackjack playing strategies:

Method	Description
Standard GP	Tree-based genetic programming with unconstrained search
Grammar GP	Grammar-guided GP using BNF constraints
True GE	Grammatical Evolution with integer genome → phenotype mapping

Evolved strategies are decision trees that make hit/stand/double decisions based on:

• Player's hand value (2-21)
• Dealer's upcard (2-11)
• Soft hand status (usable Ace)
• Double-down availability

See DESIGN.md for architecture, EA features, and design decisions.

---

Installation

Prerequisites: Python 3.8+

git clone https://github.com/oisinosullivan9/Evolutionary-Algorithms.git
cd Evolutionary-Algorithms
pip install -r requirements.txt

Optional: Install Graphviz for decision tree visualization (add to PATH during install).

---

Usage

cd project

# Standard GP
python main.py --method standard

# Grammar-Guided GP  
python main.py --method grammar

# True Grammatical Evolution
python main.py --method ge

---

Output

Files Generated

File	Description
graphs/fitness_over_generations_{method}_run{N}.png	Fitness progression plot
graphs/complexity_vs_fitness_{method}_run{N}.png	Tree size vs fitness
graphs/{method}_best_strategy_tree.png	Best strategy visualization