Genetic Algorithm Experiment

This is an experimental implementation of a genetic algorithm in Rust, created for learning and performance optimization purposes. The project demonstrates various genetic algorithm concepts while focusing on performance optimizations using parallel processing and efficient data structures.

Overview

This genetic algorithm implementation includes:

• Population initialization with random chromosomes
• Fitness calculation
• Selection (using binary search and batch processing)
• Crossover (single-point and uniform)
• Mutation
• Parallel processing using Rayon
• Performance optimizations for different chromosome sizes

Structure

The project is organized into several modules:

• main.rs: Main program logic and configuration
• frog.rs: Definition of the individual (Frog) and its traits
• crossover.rs: Crossover operations implementation
• mutation.rs: Mutation operations implementation
• select.rs: Selection algorithm implementation
• helper.rs: Utility functions for population management and I/O

Features

• Configurable population size and chromosome length
• Adjustable mutation rate
• Progress tracking with ETA and throughput calculation
• CSV file I/O for population state
• Optimized performance for both small and large chromosomes
• Multi-threaded processing using Rayon

Running the Project

1. Make sure you have Rust installed on your system.

2. Clone the repository and navigate to the project directory.

3. Build and run the project:

cargo run --release

Note: The --release flag is important for performance as it enables optimizations.

Configuration

You can modify these constants in main.rs to experiment with different parameters:

const N: usize = 12 * 26;      // Population size
const D: usize = 64;           // Chromosome size
const MUTATION_RATE: f64 = 0.001;
let generations: i32 = 100;    // Number of generations

Output

The program outputs:

• Initial population statistics
• Generation progress with throughput and ETA
• Final population statistics
• Total execution time

File I/O

The program creates two CSV files:

• initial_population.csv: Initial population state
• final_population.csv: Final population state

Note

This is an experimental project created for learning purposes and performance optimization experiments. It may not be suitable for production use without further modifications and testing.

Performance Considerations

• The implementation includes optimizations for different chromosome sizes
• Parallel processing is used where beneficial
• Memory access patterns are optimized for better cache utilization
• Batch processing is implemented for certain operations

License

Feel free to use and modify this code for your own learning and experimentation purposes.