仿生学习：智能体系计划的灵感与实现 - IT评测·应用市场-qidao123.com技术社区

import numpy as np
# 个体定义
class Individual:
def __init__(self, chromosome):
self.chromosome = chromosome
self.fitness = self.calculate_fitness()
def calculate_fitness(self):
# 自定义的适应度函数，可以根据具体应用调整
return np.sum(self.chromosome)
# 遗传算法流程
class GeneticAlgorithm:
def __init__(self, population_size, chromosome_length, generations):
self.population_size = population_size
self.chromosome_length = chromosome_length
self.generations = generations
self.population = [Individual(self.random_chromosome()) for _ in range(population_size)]
def random_chromosome(self):
# 随机生成二进制染色体
return np.random.randint(2, size=self.chromosome_length)
def selection(self):
# 选择高适应度个体
sorted_population = sorted(self.population, key=lambda x: x.fitness, reverse=True)
return sorted_population[:self.population_size // 2]
def crossover(self, parent1, parent2):
# 单点交叉
crossover_point = np.random.randint(1, self.chromosome_length - 1)
child1 = np.concatenate((parent1.chromosome[:crossover_point], parent2.chromosome[crossover_point:]))
child2 = np.concatenate((parent2.chromosome[:crossover_point], parent1.chromosome[crossover_point:]))
return Individual(child1), Individual(child2)
def mutation(self, individual):
# 基因突变
mutation_point = np.random.randint(self.chromosome_length)
individual.chromosome[mutation_point] = 1 - individual.chromosome[mutation_point]
def run(self):
for generation in range(self.generations):
selected_individuals = self.selection()
new_population = selected_individuals[:]
while len(new_population) < self.population_size:
parent1, parent2 = np.random.choice(selected_individuals, 2)
child1, child2 = self.crossover(parent1, parent2)
self.mutation(child1)
self.mutation(child2)
new_population.append(child1)
new_population.append(child2)
self.population = new_population
best_individual = max(self.population, key=lambda x: x.fitness)
print(f"Generation {generation}, Best Fitness: {best_individual.fitness}")
# 初始化遗传算法
ga = GeneticAlgorithm(population_size=100, chromosome_length=10, generations=50)
ga.run()

复制代码

class Particle:
def __init__(self, position, velocity):
self.position = position
self.velocity = velocity
self.best_position = position
self.best_fitness = self.calculate_fitness()
def calculate_fitness(self):
# 自定义的适应度函数
return np.sum(self.position**2)
class PSO:
def __init__(self, num_particles, num_dimensions, max_iter):
self.num_particles = num_particles
self.num_dimensions = num_dimensions
self.max_iter = max_iter
self.particles = [Particle(np.random.uniform(-1, 1, num_dimensions), np.random.uniform(-1, 1, num_dimensions)) for _ in range(num_particles)]
self.global_best_position = None
self.global_best_fitness = float('inf')
def update_particles(self):
for particle in self.particles:
# 更新速度和位置
inertia = 0.5
cognitive = 1.5
social = 1.5
r1, r2 = np.random.rand(), np.random.rand()
particle.velocity = (inertia * particle.velocity +
cognitive * r1 * (particle.best_position - particle.position) +
social * r2 * (self.global_best_position - particle.position))
particle.position += particle.velocity
# 更新个体和全局最佳位置
fitness = particle.calculate_fitness()
if fitness < particle.best_fitness:
particle.best_fitness = fitness
particle.best_position = particle.position
if fitness < self.global_best_fitness:
self.global_best_fitness = fitness
self.global_best_position = particle.position
def run(self):
for iteration in range(self.max_iter):
self.update_particles()
print(f"Iteration {iteration}, Best Fitness: {self.global_best_fitness}")
# 初始化PSO
pso = PSO(num_particles=30, num_dimensions=5, max_iter=100)
pso.run()

复制代码