人工智能实验：王浩算法python实现（附算法设计图）
人工智能王浩算法python实现（附算法设计图）
实验课作业，没什么实际价值，不过还挺难的，copy一下应付作业没问题。
原作者： 王浩算法—DMU.
他的版本写的比较晦涩，我改了很多，加入了与、或、等价的处理方法，效率提高不少，当然也容易理解了，顺便把算法设计图写出来了。
一、算法设计图：
二、完整代码
验证案例：
s = ‘=> a -> ( b -> a )’
s = “=> ( a -> b ) -> ( ( ! b ) -> ( ! a ) )”
s = ‘=> ( a -> b ) -> ( ( a -> c ) -> ( a -> ( b & c ) ) )’
s = ‘=> ( a -> c ) -> ( ( b -> c ) -> ( ( a | b ) -> c ) )’
#s = ‘=> ( a <-> b ) -> ( a -> b )’
注：实验环境python3.8，pycharm
直接跑就行，在main函数那里可以改问题输入
import copy
# 列表转字符串
def listtostr(m):
s = ''
for i in m:
s = s + i + ' '
return s
# 识别最外层括号，括号位置索引
def braces(m):
o = []
p = []
bralist = []
for i in range(len(m)):
if m[i] == '(':
o.append(-1)
p.append(i)
# print('i',i,m[i])
elif m[i] == ')':
o.append(1)
p.append(i)
# print('i',i,m[i])
else:
continue
t = sum(o)
if t == 0:
bralist.append([p[0], p[-1]])
o = []
p = []
# print('最外层括号:',bralist)
return bralist
# 识别最外层逗号(相继式符号)，返回最外层逗号的位置列表
def comma(m):
comlist = [-1] # 为字符串开头加一标志位
for i in range(len(m)):
if m[i] == ',' or m[i] == '=>': # 逗号和相继式符号位置添加到逗号列表
comlist.append(i)
return comlist
# 识别符号！和->
def detect(dm):
dmkuo = braces(dm)
if dm[0] == '!':
return 1
elif dm[1] == '->':
return 2
elif dm[1] == '&':
return 3
elif dm[1] == '|':
return 4
elif dm[1] == '<->':
return 5
elif dmkuo!=[]:
if dm[dmkuo[0][-1] + 1] == '->':
return 2
elif dm[dmkuo[0][-1] + 1] == '&':
return 3
elif dm[dmkuo[0][-1] + 1] == '|':
return 4
elif dm[dmkuo[0][-1] + 1] == '<->':
return 5
#找出需要处理的字段
def find_(m, comlist):
# print('find开始：-----------')
t = m.index('=>')
if comlist[1] - 1 == comlist[0]:#删除在开头的=>
comlist.pop(0)
k = 0
for i in comlist:
if i + 2 == len(m) or m[i + 2] == ',' or m[i + 2] == '=>':#长度小于2的单一字段不需要处理
# print('单一元素位置：\n',m[i+1:i+2],i+1)
k = k + 1
continue
if comlist[k] == comlist[-1]:#仅剩一个逗号的情况，处理后半段
last = len(m)
else:
last = comlist[k + 1]
dm = m[i + 1:last]
print('要处理的子段是：', listtostr(dm))
dou = last
j = detect(dm)
if j == 1:
if i < t:
return nonleft(dm, m, dou - 1, i)
elif i >= t:
return nonright(dm, m, dou - 1, i)
break
elif j == 2:
if i < t:
return yunleft(dm, m, dou , i)
elif i >= t:
return yunright(dm, m, dou , i)
break
elif j == 3:
return hequ(dm , m , dou , i)
elif j == 4:
return xiqu(dm , m , dou , i)
elif j == 5:
return dengjia(dm , m , dou , i)
k = k + 1
# 处理多余逗号
def dd(m):
for i in range(len(m)):
if i == len(m) - 1:
if m[i] == ',':
m.pop()
break
elif m[i] == ',' and m[i + 1] == ',':
m.pop(i)
elif m[i] == ',' and m[i + 1] == '=>':
m.pop(i)
elif m[i] == '=>' and m[i + 1] == ',':
m.pop(i+1)
return m
#i:开始位置 dou:结束位置 ss：要处理的字段
# 按照四种情况处理 M 列表
# 若 a ,!X,b => r 则 a, b =>X,r
def nonleft(ss, m, dou, i):
ss.pop(0)#删除非
kuoss = braces(ss)
if kuoss != []:#删除括号
ss.pop(0)
ss.pop(-1)
for x in range(dou, i, -1):#删除非字段
m.pop(x)
m = dd(m)#删除多余逗号
t = m.index('=>')
ss.append(",")
ss.insert(0,"=>")
m[t:t+1]=ss
# print('\n"!"在"=>"左边\n', '-------m-------\n',listtostr(m))
return [[m, '规则1', ' ', [0]]]
# 若 a ,b =>!X, r 则 a, X =>b , r
def nonright(ss, m, dou, i):
ss.pop(0)#删除非
kuoss = braces(ss)
if kuoss != []:#删除括号
ss.pop(0)
ss.pop(-1)
for x in range(dou, i, -1):#删除非字段
m.pop(x)
m = dd(m)#删除多余逗号
t = m.index('=>')
if t==0:
ss.append("=>")
m[t:t+1]=ss
else:
ss.insert(0, ",")
ss.append("=>")
m[t:t+1]=ss
# print('\n"!"在"=>"右边\n','-------m-------\n',listtostr(m))
return [[m, '规则2', ' ', [0]]]
#找出最外层需要处理的->位置
def yunhan(ss):
dmkuo = braces(ss)
if ss[1] == '->':
return 1
else:
return dmkuo[0][-1] + 1
# 若 a ,(X -> Y),b => r 则 a ,!X,b => r ; a ,Y,b => r
def yunleft(ss, m, dou, i):
yun = yunhan(ss)
ssx = ss[:yun]
ssy = ss[yun + 1:]
kuossy = braces(ssy)
if kuossy != []:
ssy.pop(0)
ssy.pop(-1)
ssx.insert(0, '!')
#m = deal(dou, i)
m1 = copy.deepcopy(m)
m2 = copy.deepcopy(m)
if i==0:
i=i-1
m1[i+1:dou]=ssx
m2[i+1:dou]=ssy
# print('\n"->"在"=>"左边\n',\
# '--------m1-------\n',listtostr(m1),\
# '\n--------m2-------\n',listtostr(m2))
# return listtostr(m1),listtostr(m2)
return [[m1, '规则3', ' ', [0]], [m2, '规则3', ' ', [0]]]
# 若 a => b, (X -> Y), r 则 a => b, ！X ， Y , r
def yunright(ss, m, dou, i):
yun = yunhan(ss)#找出最外层—>位置
ssx = ss[:yun]
ssy = ss[yun + 1:]
kuossy = braces(ssy)
if kuossy != []:
ssy.pop(0)
ssy.pop(-1)
ssx.insert(0, '!')
ssxy = ssx + [','] + ssy
m[i+1:dou]=ssxy
# print('\n"->"在"=>"右边\n','-------m-------\n',listtostr(m))
return [[m, '规则4', ' ', [0]]]
#对等价，合取，等价进行转换
#p&q to !(p->!q) p|q to !p->q p<->q to !((p->q)->(!(q->p)))
def exchange(m):
if '&' in m:
t=m.index("&")
p=m[0:t]
q=m[t+1:len(m)]
return ["!","("]+p+["->"]+q+[")"]
elif '|' in m:
t = m.index("|")
p = m[0:t]
q = m[t + 1:len(m)]
return ["!"] + p + ["->"] + q
elif '<->' in m:
t = m.index("<->")
p = m[0:t]
q = m[t + 1:len(m)]
return ["!","(","("] + p + ["->"] + q+[")","->","(","!","("]+ p + ["->"] + q+[")",")",")"]
def hequ(ss, m, dou, i):
ss=exchange(ss)
m[i + 1:dou]=ss
return [[m, '合取符号转换', ' ', [0]]]
def xiqu(ss, m, dou, i):
ss = exchange(ss)
m[i + 1:dou] = ss
return [[m, '析取符号转换', ' ', [0]]]
def dengjia(ss, m, dou, i):
ss = exchange(ss)
m[i + 1:dou] = ss
return [[m, '等价符号转换', ' ', [0]]]
if __name__=='__main__':
s = '=> a -> ( b -> a )'
s = "=> ( a -> b ) -> ( ( ! b ) -> ( ! a ) )"
s = '=> ( a -> b ) -> ( ( a -> c ) -> ( a -> ( b & c ) ) )'
s = '=> ( a -> c ) -> ( ( b -> c ) -> ( ( a | b ) -> c ) )'
#s = '=> ( a <-> b ) -> ( a -> b )'
# s = input('请输入要推理的字符串(字符之间用空格做间隔)：\n')
# s = '=>'+' '+s
m = s.split()
e = 0
stock = []
stock.append([m, '结果', ' ', [0]])
stock1 = []
# print('first stock',stock)
while (stock):
e = e + 1
y = 0
print('\t\t*****', e, '*****')
mm = stock.pop()
# print("输出",mm)
m = mm[0]
if '->' not in m and '!' not in m and '|' not in m and '&' not in m and '<->' not in m :
d = m.index('=>')
m1 = m[:d]
m2 = m[d + 1:]
for i in m1: # 有相同元素，是公理
if i != ',':
if i in m2:
print(listtostr(m), '公理', '\n -------END------')
stock1.append([listtostr(m), mm[1], '公理', mm[3]])
break
else:
y = y + 1
for i in m1:
if i == ',':
m1.remove(i)
if y == len(m1):
print(listtostr(m), '不是公理', '\n -------END------')
stock1.append([listtostr(m), mm[1], '不是公理', mm[3]])
break
continue
stock1.append([listtostr(m), mm[1], mm[2], mm[3]])
print('要处理的字符串:', listtostr(m))
comlist = comma(m)
f = find_(m, comlist)
print('处理后的结果:')
for i in range(len(f)): # 规则三返回两个结果
f[i][3] = [mm[3][0] + i]
# f[i][3].append(mm[3][i]+i)
print(listtostr(f[i][0]), f[i][1], '\n -------END------', )
stock.append(f[i])
for i in stock1:
if i[2] == '不是公理':
print("推理失败!!!\t" + i[0] + i[2])
exit()
# 获取规则三的来源
'''
for i in stock1:
print(i)
'''
k = []
for i in range(len(stock1)):
k.append([len(stock1) - i, stock1[i][3][0]])
rule3 = []
for x in range(len(k) - 1):
if k[x][1] == k[x + 1][1] - 1:
for i in range(x + 1, len(k)):
if k[i][1] == k[x][1]:
index = i
rule3.append([k[x][0], k[index][0], k[x + 1][0]])
break
for i in k:
print(i)
#print(rule3)
stock1.reverse()
print("__________")
for x in stock1:
print(x)
p = []
for i in range(len(stock1)):
flag = 1
for x in rule3:
if i + 1 == x[0]:
p.append([stock1[i][0], "由[" + str(x[1]) + "]、[" + str(x[2]) + "]和" + stock1[i - 1][1]])
flag = 0
if flag == 1:
if stock1[i][2] == '公理':
p.append([stock1[i][0], "公理"])
else:
p.append([stock1[i][0], "由[" + str(i) + "]和" + stock1[i - 1][1]])
print("推理过程：")
print("_" * 100)
for i in range(len(p)):
print("[" + str(i + 1) + "]", "{:<60}".format(p[i][0]), "\t\t", p[i][1])
print("推理结束")