P:S:我从this元线程中了解到,Stack Overflow也可以提出代码解释。
我试图从this网站了解Aristotle Number拼图的解算器。我理解这一点,直到我们使用高斯消除进行行减少并找到以下内容:
a = 76 - j - k - n - 2o - p - r - s
b = j + n + o
c = -38 + k + o + p + r + s
d = j + k + o
e = -38 + k + n + o + p + r
f = 38 - j - k - n - o - p
g = 38 - k - o - r
h = -38 + n + o + p + r + s
i = 38 - j - k - n - o - r
l = 38 - p - s
m = 38 - n - o - p
q = 38 - r - s
然后代码的作者继续说道:
现在,从{1,2,...,19}获取大小为7的每个排列,将其分配给自变量,生成从属变量并对约束进行测试,直到找到解。
我真的不理解这个概念。特别是在this c文件中,我不了解以下两个函数:
bool next_permutation(void)
{
for (int x = 6; x >= 0; x--) {
indices[x]++;
if (indices[x] == 19) {
if (!x) {
return false;
}
moveback(x, 18);
indices[x] = x;
continue;
}
swap(x, indices[x]);
break;
}
j = elem[0];
k = elem[1];
n = elem[2];
o = elem[3];
p = elem[4];
r = elem[5];
s = elem[6];
return true;
}
// adds values to set
// returns true if value successfully added; false otherwise
bool add(int value)
{
if (value > 19 || value < 1) {
return false;
}
int bit = 1 << value;
if (set & bit) {
return false;
}
set |= bit;
return true;
}
如果有人能帮我理解这个解算器,我将非常感激。请注意,作者使用python脚本进行行减少。
答案 0 :(得分:0)
我认为Gaussian elimination只是一种通过将一个方程代入其他方程来消除不必要的变量来减少一组方程的方法。例如,您可以采用19个方程中的第一个,并将其转换为以下公式:
a = 38 - b - c
然后将其替换为另一个18.这将消除a并将你的等式减少到18.然后重复直到你不能再消除它。
这是一个很好的问题,它给了我一个尝试SymPy包的借口。
这里有一些代码可以在SymPy中创建方程式,然后使用SymPy的solve
函数来减少方程式。
from sympy import symbols, solve
from sympy.parsing.sympy_parser import parse_expr
# Number of unknowns
n = 19
variable_names = [chr(c + 97) for c in range(19)]
# Create SymPy variables
variables = symbols(variable_names)
print("\n{} unknowns:".format(len(variables)))
print(variables)
# These strings define the equations to be solved
hexagon_rows = [
'abc',
'defg',
'hijkl',
'mnop',
'qrs',
'cgl',
'bfkp',
'aejos',
'dinr',
'hmq',
'lps',
'gkor',
'cfjnq',
'beim',
'adh'
]
def make_expression(chars, rhs=0):
return parse_expr('+'.join(list(chars)) + '-' + str(rhs))
expressions = []
for chars in hexagon_rows:
expressions.append(make_expression(chars, 38))
print("\n{} equations to solve:".format(len(expressions)))
for expr in expressions:
print("{} = 0".format(expr))
# Try to solve the equations
# (They can't be solved but SymPy reduces them down)
reduced_expressions = solve(expressions)
print("\nReduced to {} equations:".format(len(reduced_expressions)))
for var, expr in reduced_expressions.items():
print("{} = {}".format(var, expr))
输出:
19 unknowns:
[a, b, c, d, e, f, g, h, i, j, k, l, m, n, o, p, q, r, s]
15 equations to solve:
a + b + c - 38 = 0
d + e + f + g - 38 = 0
h + i + j + k + l - 38 = 0
m + n + o + p - 38 = 0
q + r + s - 38 = 0
c + g + l - 38 = 0
b + f + k + p - 38 = 0
a + e + j + o + s - 38 = 0
d + i + n + r - 38 = 0
h + m + q - 38 = 0
l + p + s - 38 = 0
g + k + o + r - 38 = 0
c + f + j + n + q - 38 = 0
b + e + i + m - 38 = 0
a + d + h - 38 = 0
Reduced to 12 equations:
j = b - n - o
k = e - n - o - p - r + 38
l = -p - s + 38
i = -b - e + n + o + p
c = e - n + s
f = -b - e + n + o + r
g = -e + n + p
q = -r - s + 38
m = -n - o - p + 38
h = n + o + p + r + s - 38
d = b + e - 2*n - o - p - r + 38
a = -b - e + n - s + 38
希望有所帮助。至少第一部分!