蚁群算法

蚁群算法的Python代码及其效果演示（含注释）

以下为基本蚁群算法的Python代码（含注释）。

随时可以运行：

fromturtleimport*

fromrandomimport*

fromjsonimportload

k=load(open(""))

city_num,ant_num=30,30#规定城市和蚂蚁总数

x_data=k[0]#城市的x坐标之集合

y_data=k[1]#城市的y坐标之集合

best_length=float("inf")

best_path=[]

alpha=1

beta=7

rho=0.5

potency_list=[1forxxinrange(city_num**2)]

Q=1

##城市的index从0开始

#下面列表存储城市间距离

defget_i_index(n):

ifn%city_num==0:

returnn//city_num-1

el:

returnn//city_num

defget_j_index(n):

ifn%city_num==0:

returncity_num-1

el:

returnn%city_num-1

distance_list=[((x_data[get_i_index(z)]-x_data[get_j_index(z)])**2+(y_data[get_i_index(z)]-

y_data[get_j_index(z)])**2)**0.5forzinrange(1,city_num**2+1)]

classant(object):

def__init__(lf,ant_index):

_index=ant_index

=list(range(city_num))

t_length=0

t_city=randint(0,city_num-1)

l_city=t_city

(t_city)

=[t_city]

=0

#根据城市的index求出两城市间距离

defget_distance(lf,index_1,index_2):

returndistance_list[index_1*city_num+index_2]

defget_potency(lf,index_1,index_2):

returnpotency_list[index_1*city_num+index_2]

defget_prob_list(lf):

res=[_potency(t_city,x)**alpha*(1/_distance(t_city,x))**bet

]

sum_=sum(res)

final_res=[y/sum_foryinres]

returnfinal_res

##轮盘赌选择城市

def__choo_next_city(lf):

city_list=

prob_list=_prob_list()

tmp=random()

sum_=0

forcity,probinzip(city_list,prob_list):

sum_+=prob

ifsum_>=tmp:

+=_distance(t_city,city)

t_city=city

(t_city)

(t_city)

return

defrunning(lf):

globalbest_length,best_path

forxinrange(city_num-1):

lf.__choo_next_city()

+=_distance(t_city,l_city)

(l_city)

best_length=

best_path=

return(,)

defgo():

operation=[]

forxinpotency_list:

x*=(1-rho)

forxinrange(ant_num):

(ant(x).running())

forxinoperation:

foryinrange(city_num-1):

potency_list[x[0][y]*city_num+x[0][y+1]]+=Q/x[1]

#print(f"potency_list:{potency_list}")

#print(f"best_path:{best_path}")

#print(f"best_length:{best_length}")

foryyinrange(1000):

go()

print(f"best_length:{best_length}")

pu()

tpos(x_data[best_path[0]],y_data[best_path[0]])

pd()

forxinrange(1,city_num+1):

tpos(x_data[best_path[x]],y_data[best_path[x]])

运行效果：