Genetic Algorithm Pathfinding

This repository contains a Python implementation of a genetic algorithm to solve a pathfinding problem on a 2D grid. The goal is to navigate from starting points on the grid to maximize a score based on specific rules. This is actually a tool I used to help me play the Muterra Discord Game - it was by no means a winner, but was definitely a time saver and fun project

Code Overview

mutate_agent(agent_genome)

Mutates a given agent's genome by randomly altering its directions.

compress(solution)

Compresses a solution by removing unnecessary moves.

generate_random_agent()

Generates a random agent with a valid path from a random starting point.

pretty_print(solution)

Prints the solution in a readable format with the path and score.

compute_fitness(solution)

Calculates the fitness of a given solution based on the grid rules and path.

freeze(d) and unfreeze(d)

Utility functions to handle immutable and mutable data structures for the genetic algorithm.

generate_random_population(pop_size)

Generates a random population of agents.

run_genetic_algorithm(generations=50000, population_size=100)

Runs the genetic algorithm for a specified number of generations and population size.

Getting Started

Prerequisites

• Python 3.x
• Required libraries: datetime, copy, json, sys, random, os, base64, hashlib

Running the Code

1. Clone the repository:

   git clone https://github.com/yourusername/genetic-algorithm-pathfinding.git
   cd genetic-algorithm-pathfinding

2. Run the script:

   python pathfinding.py

Example Output

The script will output the generation number, best fitness score, unique genomes, and the path with its score.

Generation 1000 best: 55.5 | Unique genomes: 1000 [10.123]
[A] N-S-E-W-N-...