genetic-algorithms

## What I do - Implement evolutionary algorithms - Design genetic representations - Create fitness functions - Implement selection and crossover - Handle mutation operators - Optimize using evolution ## When to use me Use me when: - Complex optimization problems - Noisy or non-differentiable fitness - Multi-objective optimization - Feature selection - Neural architecture search ## Key Concepts ### Genetic Algorithm Flow ``` ┌──────────────┐ │ Population │◀──────────────┐ │ Generation │ │ └──────┬───────┘ │ │ │ ▼ │ ┌──────────────┐ ┌─────────┐│ │ Evaluate │ │ Select ││ │ Fitness │───▶│ Parents││ └──────────────┘ └────┬────┘ │ ▼ ┌─────────────┐ │ Crossover │ │ + Mutate │ └──────┬──────┘ │ ▼ ┌─────────────┐ │ Replace │ │ Population │ └──────┬──────┘ │ ▼ ┌──────────────┐ │ Continue? │ └──────────────┘ ``` ### Implementation ```python import random import numpy as np class GeneticAlgorithm: def __init__(self, pop_size=100, mutation_rate=0.1, crossover_rate=0.8): self.pop_size = pop_size self.mutation_rate = mutation_rate self.crossover_rate = crossover_rate def init_population(self): return [self.create_individual() for _ in range(self.pop_size)] def fitness(self, individual): raise NotImplemented def selection(self, population, fitnesses): # Tournament selection selected = [] for _ in range(len(population)): i, j = random.sample(range(len(population)), 2) if fitnesses[i] > fitnesses[j]: selected.append(population[i].copy()) else: selected.append(population[j].copy()) return selected def crossover(self, parent1, parent2): if random.random() > self.crossover_rate: return parent1.copy(), parent2.copy() point = random.randint(1, len(parent1) - 1) child1 = parent1[:point] + parent2[point:] child2 = parent2[:point] + parent1[point:] return child1, child2 def mutate(self, individual): for i in range(len(individual)): if random.random() < self.mutation_rate: individual[i] = self.random_gene() return individual def evolve(self, generations=100): population = self.init_population() for gen in range(generations): fitnesses = [self.fitness(ind) for ind in population] # Elitism - keep best best_idx = np.argmax(fitnesses) new_pop = [population[best_idx].copy()] # Selection parents = self.selection(population, fitnesses) # Crossover and mutation for i in range(0, len(parents) - 1, 2): child1, child2 = self.crossover(parents[i], parents[i+1]) new_pop.append(self.mutate(child1)) new_pop.append(self.mutate(child2)) population = new_pop[:self.pop_size] return max(population, key=self.fitness) ``` ### Operators - **Selection**: Tournament, roulette, rank - **Crossover**: One-point, two-point, uniform - **Mutation**: Bit flip, swap, Gaussian - **Replacement**: Generational, elitism