【实例简介】花授粉算法.m
【实例截图】

【核心代码】
importnumpy as np
frommatplotlibimportpyplot
as
plt
importrandom
# 初始化种群
definit(n_pop, lb, ub, nd):
"""
:param n_pop: 种群
:param lb: 下界
:param ub: 上界
:param nd: 维数
"""
p=lb (ub-lb)*np.random.rand(n_pop, nd)
returnp
# 适应度函数
defsphere(x):
y=np.sum(x**2,1)
returny
defAckley_1(x):
n, d=x.shape
y=-20*np.exp(-0.02*np.sqrt(1/d*np.sum(x**2,1)))-np.exp(
1/d*np.sum(np.cos(2*np.pi*x),1)) 20 np.e
returny
defAckley_2(x):
y=-200*np.exp(-0.02*np.sqrt(x[:,0]**2 x[:,1]**2))
returny
defAckley_3(x):
y=-200*np.exp(-0.02*np.sqrt(x[:,0]**2 x[:,1]**2)) 5*np.exp(
np.cos(3*x[:,0]) np.sin(3*x[:,1]))
returny
defAckley_4(x, y=0):
_, d=x.shape
foriinrange(1, d):
y =np.exp(-0.2*np.sqrt(x[:, i-1]**2 x[:, i]**2)) 3*(
np.cos(2*x[:, i-1]) np.sin(2*x[:, i]))
returny
defAdjiman(x):
y=np.cos(x[:,0])*np.sin(x[:,1])-x[:,0]/(x[:,1]**2 1)
returny
defAlpine(x):
y=np.sum(np.abs(x*np.sin(x) 0.1*x),1)
returny
defAlpine2(x):
y=np.prod(np.sqrt(x)*np.sin(x), axis=1)
returny
defBartels(x):
y=np.abs(x[:,0]**2 x[:,1]**2 x[:,0]*x[:,1]) np.abs(np.sin(x[:,0])) np.abs(np.c
returny
defBeale(x):
y=(1.5-x[:,0] x[:,0]*x[:,1])**2 (2.25-x[:,0] x[:,0]*x[:,1]**2)**2 (
2.625-x[:,0] x[:,0]*x[:,1]**3)**2
returny
f_score=sphere # 函数句柄
# Levy飞行Beale
defLevy(nd, beta=1.5):
num=np.random.gamma(1 beta)*np.sin(np.pi*beta/2)
den=np.random.gamma((1 beta)/2)*beta*2**((beta-1)/2)
sigma_u=(num/den)**(1/beta)
u=np.random.normal(0, sigma_u**2, (1, nd))
v=np.random.normal(0,1, (1, nd))
z=u/(np.abs(v)**(1/beta))
returnz
defFPA(Max_g, n_pop, Pop, nd, lb, ub, detail): # FPA算法
"""
:param Max_g: 迭代次数
:param n_pop: 种群数目
:param Pop: 花粉配子
:param nd: 维数
:param lb: 下界
:param ub: 上界
:param detail: 显示详细信息
"""
# 计算初始种群中最好个体适应度值
pop_score=f_score(Pop)
g_best=np.min(pop_score)
g_best_loc=np.argmin(pop_score)
g_best_p=Pop[g_best_loc, :].copy()
# 问题设置
p=0.8
best_fit=np.empty((Max_g,))
# 迭代
foritinrange(1, Max_g 1):
foriinrange(n_pop):
ifnp.random.rand() < p:
new_pop=Pop[i, :] Levy(nd)*(g_best_p-Pop[i, :])
new_pop=np.clip(new_pop, lb, ub) # 越界处理
else:
idx=random.sample(list(range(n_pop)),2)
new_pop=Pop[i, :] np.random.rand()*(Pop[idx[1], :]-Pop[idx[0], :])
new_pop=np.clip(new_pop, lb, ub) # 越界处理
iff_score(new_pop.reshape((1,-1))) < f_score(Pop[i, :].reshape((1,-1))):
Pop[i, :]=new_pop
# 计算更新后种群中最好个体适应度值
pop_score=f_score(Pop)
new_g_best=np.min(pop_score)
new_g_best_loc=np.argmin(pop_score)
ifnew_g_best < g_best:
g_best=new_g_best
g_best_p=Pop[new_g_best_loc, :].copy()
best_fit[it-1]=g_best
ifdetail:
print("----------------{}/{}--------------".format(it, Max_g))
print(g_best)
print(g_best_p)
returnbest_fit, g_best
if__name__=="__main__":
pop=init(30,-100,100,2)
fitness, g_best=FPA(1000,30, pop,2,-100,100,True)
# 可视化
plt.figure()
# plt.plot(fitness)
plt.semilogy(fitness)
# 可视化
# fig = plt.figure()
# plt.plot(p1, fit)
plt.show()
网友评论
我要评论