fix: 统一文件编码为 utf-8 并将 benchmark 日志改为英文
This commit is contained in:
169
solutions/0066.DiophantineEq/README.md
Normal file
169
solutions/0066.DiophantineEq/README.md
Normal file
@@ -0,0 +1,169 @@
|
||||
# 佩尔方程(Pell's Equation)—— 基本解的完整数学理论
|
||||
|
||||
## 一、方程定义与历史背景
|
||||
|
||||
**佩尔方程**是指一类特殊的二元二次不定方程:
|
||||
|
||||
$$x^2 - Dy^2 = 1$$
|
||||
|
||||
其中 $D$ 是一个**非平方正整数**(若 $D$ 为完全平方数,则方程退化为 $(x - y\sqrt{D})(x + y\sqrt{D}) = 1$,仅有平凡整数解)。
|
||||
|
||||
### 历史脉络
|
||||
- **公元前400年左右**:古希腊数学家就已研究 $x^2 - 2y^2 = 1$,发现解 $(3,2), (17,12), (99,70), \ldots$
|
||||
- **Brahmagupta(628年)**:印度数学家系统研究了"Varga-prakriti"方程,提出"bhāvanā"合成法则,本质上发现了 $(x_1 + y_1\sqrt{D})^n$ 的通解结构
|
||||
- **Bhaskara II(1150年)**:在《Līlāvatī》中给出求解 $x^2 - 61y^2 = 1$ 的算法(循环法 cakravāla),得到基本解 $(1766319049, 226153980)$
|
||||
- **Fermat(1657年)**:向欧洲数学界挑战求解此方程,Euler 误将解法归功于 Pell,由此得名
|
||||
|
||||
---
|
||||
|
||||
## 二、基本理论:解的存在性与结构
|
||||
|
||||
### 定理 1(Lagrange, 1768):解的存在性
|
||||
对任意非平方正整数 $D$,方程 $x^2 - Dy^2 = 1$ **总有无穷多组正整数解**。
|
||||
|
||||
### 定义:基本解(最小解)
|
||||
所有正整数解中使 $x + y\sqrt{D}$ 最小的那组解 $(x_1, y_1)$ 称为**基本解**(或**最小解**、**本源解**)。它是生成全部解的"种子"。
|
||||
|
||||
### 定理 2:通解结构
|
||||
若 $(x_1, y_1)$ 是基本解,则**所有**正整数解 $(x_n, y_n)$ 满足:
|
||||
|
||||
$$x_n + y_n\sqrt{D} = (x_1 + y_1\sqrt{D})^n, \quad n = 1, 2, 3, \ldots$$
|
||||
|
||||
等价地,解可通过递推得到:
|
||||
$$\begin{cases} x_{n+1} = x_1 x_n + D y_1 y_n \\ y_{n+1} = y_1 x_n + x_1 y_n \end{cases}$$
|
||||
|
||||
或写成矩阵形式:
|
||||
$$\begin{pmatrix} x_{n+1} \\ y_{n+1} \end{pmatrix} = \begin{pmatrix} x_1 & Dy_1 \\ y_1 & x_1 \end{pmatrix} \begin{pmatrix} x_n \\ y_n \end{pmatrix}$$
|
||||
|
||||
---
|
||||
|
||||
## 三、核心算法:连分数法
|
||||
|
||||
求解基本解的标准方法是**连分数展开** $\sqrt{D}$。
|
||||
|
||||
### 3.1 $\sqrt{D}$ 的连分数展开
|
||||
|
||||
对非平方正整数 $D$,$\sqrt{D}$ 有**周期性连分数展开**:
|
||||
|
||||
$$\sqrt{D} = [a_0; \overline{a_1, a_2, \ldots, a_L}]$$
|
||||
|
||||
其中 $a_0 = \lfloor\sqrt{D}\rfloor$,$(a_1, a_2, \ldots, a_L)$ 是纯循环部分,$L$ 为周期长度。
|
||||
|
||||
### 3.2 计算递推关系
|
||||
|
||||
展开算法的核心递推(计算周期中各项):
|
||||
$$\begin{aligned}
|
||||
m_0 &= 0, \quad d_0 = 1, \quad a_0 = \lfloor\sqrt{D}\rfloor \\
|
||||
m_{k+1} &= d_k \cdot a_k - m_k \\
|
||||
d_{k+1} &= \frac{D - m_{k+1}^2}{d_k} \\
|
||||
a_{k+1} &= \left\lfloor \frac{a_0 + m_{k+1}}{d_{k+1}} \right\rfloor
|
||||
\end{aligned}$$
|
||||
|
||||
当 $(m_k, d_k, a_k)$ 首次重复时,即得到一个完整周期。
|
||||
|
||||
### 3.3 渐近分数(Convergents)
|
||||
|
||||
记第 $n$ 个渐近分数为 $\frac{p_n}{q_n}$,递推公式:
|
||||
$$\begin{cases}
|
||||
p_{-1} = 1, & p_0 = a_0, \quad p_n = a_n p_{n-1} + p_{n-2} \\[4pt]
|
||||
q_{-1} = 0, & q_0 = 1, \quad q_n = a_n q_{n-1} + q_{n-2}
|
||||
\end{cases}$$
|
||||
|
||||
### 3.4 关键定理(Lagrange)
|
||||
|
||||
> **周期 $L$ 为偶数**:基本解 $(x_1, y_1) = (p_{L-1}, q_{L-1})$,即第 $(L-1)$ 个渐近分数满足 $p_{L-1}^2 - D q_{L-1}^2 = 1$
|
||||
>
|
||||
> **周期 $L$ 为奇数**:基本解 $(x_1, y_1) = (p_{2L-1}, q_{2L-1})$,需要走到第 $(2L-1)$ 个渐近分数
|
||||
|
||||
---
|
||||
|
||||
## 四、经典算例
|
||||
|
||||
### 例 1:$D = 2$
|
||||
|
||||
$$\sqrt{2} = [1; \overline{2}], \quad L = 1 \text{(奇数)}$$
|
||||
|
||||
渐近分数序列:
|
||||
$$\frac{1}{1}, \frac{3}{2}, \frac{7}{5}, \frac{17}{12}, \ldots$$
|
||||
|
||||
检验:$3^2 - 2 \cdot 2^2 = 9 - 8 = 1$ ✓
|
||||
|
||||
**基本解:$(x_1, y_1) = (3, 2)$**
|
||||
|
||||
通解:$(3 + 2\sqrt{2})^n = x_n + y_n\sqrt{2}$,前几项:
|
||||
- $n=2$: $(17, 12)$,$17^2 - 2 \cdot 12^2 = 289 - 288 = 1$
|
||||
- $n=3$: $(99, 70)$,$99^2 - 2 \cdot 70^2 = 9801 - 9800 = 1$
|
||||
- $n=4$: $(577, 408)$
|
||||
|
||||
### 例 2:$D = 13$
|
||||
|
||||
$$\sqrt{13} = [3; \overline{1, 1, 1, 1, 6}], \quad L = 5 \text{(奇数)}$$
|
||||
|
||||
需要计算到第 $2L - 1 = 9$ 个渐近分数:
|
||||
|
||||
$$\frac{p_9}{q_9} = \frac{649}{180}$$
|
||||
|
||||
验证:$649^2 - 13 \cdot 180^2 = 421201 - 421200 = 1$ ✓
|
||||
|
||||
**基本解:$(649, 180)$**
|
||||
|
||||
### 例 3:$D = 61$(历史上最著名的难题)
|
||||
|
||||
$$\sqrt{61} = [7; \overline{1, 4, 3, 1, 2, 2, 1, 3, 4, 1, 14}], \quad L = 11 \text{(奇数)}$$
|
||||
|
||||
周期长达 11,需走到第 $2 \times 11 - 1 = 21$ 个渐近分数:
|
||||
|
||||
**基本解:$(1766319049, 226153980)$**
|
||||
|
||||
验证:$1766319049^2 - 61 \times 226153980^2 = 1$ ✓
|
||||
|
||||
这正是 Bhaskara II 在 12 世纪手工计算出的惊人结果。
|
||||
|
||||
---
|
||||
|
||||
## 五、负佩尔方程:$x^2 - Dy^2 = -1$
|
||||
|
||||
### 定理 3:可解性判据
|
||||
方程 $x^2 - Dy^2 = -1$ 有整数解**当且仅当** $\sqrt{D}$ 的连分数周期 $L$ 为**奇数**。
|
||||
|
||||
### 定理 4:基本解位置
|
||||
当 $L$ 为奇数时,$x^2 - Dy^2 = -1$ 的基本解由第 $(L-1)$ 个渐近分数给出;而 $x^2 - Dy^2 = 1$ 的基本解则由第 $(2L-1)$ 个渐近分数给出。
|
||||
|
||||
### 例:$D = 5$
|
||||
|
||||
$$\sqrt{5} = [2; \overline{4}], \quad L = 1 \text{(奇数)}$$
|
||||
|
||||
- $x^2 - 5y^2 = -1$ 的基本解:$(p_0, q_0) = (2, 1)$,$2^2 - 5 \cdot 1^2 = 4 - 5 = -1$ ✓
|
||||
- $x^2 - 5y^2 = 1$ 的基本解:$(p_1, q_1) = (9, 4)$,$9^2 - 5 \cdot 4^2 = 81 - 80 = 1$ ✓
|
||||
|
||||
对比 $D = 3$($L = 2$ 偶数):$x^2 - 3y^2 = -1$ **无解**。
|
||||
|
||||
---
|
||||
|
||||
## 六、数值验证汇总
|
||||
|
||||
| $D$ | 周期 $L$ | 周期奇偶 | $x^2 - Dy^2 = 1$ 基本解 | $x^2 - Dy^2 = -1$ |
|
||||
|:---:|:---:|:---:|:---|:---|
|
||||
| 2 | 1 | 奇 | $(3, 2)$ | $(1, 1)$ ✓ |
|
||||
| 3 | 2 | 偶 | $(2, 1)$ | 无解 |
|
||||
| 5 | 1 | 奇 | $(9, 4)$ | $(2, 1)$ ✓ |
|
||||
| 6 | 2 | 偶 | $(5, 2)$ | 无解 |
|
||||
| 7 | 4 | 偶 | $(8, 3)$ | 无解 |
|
||||
| 10 | 1 | 奇 | $(19, 6)$ | $(3, 1)$ ✓ |
|
||||
| 13 | 5 | 奇 | $(649, 180)$ | $(18, 5)$ ✓ |
|
||||
| 29 | 5 | 奇 | $(9801, 1820)$ | $(70, 13)$ ✓ |
|
||||
| 61 | 11 | 奇 | $(1766319049, 226153980)$ | $(29718, 3805)$ ✓ |
|
||||
|
||||
---
|
||||
|
||||
## 七、算法复杂度与注意事项
|
||||
|
||||
1. **基本解的大小**:$(x_1, y_1)$ 可以**极其巨大**。例如 $D = 991$ 时,$x_1 \approx 3.7 \times 10^{47}$,$y_1 \approx 1.2 \times 10^{46}$。基本解的位数关于 $D$ 可以是指数级增长。
|
||||
|
||||
2. **计算精度**:实现时必须使用任意精度整数运算(如 Python 的 `int` 类型),标准 64 位整数远不够。
|
||||
|
||||
3. **与代数数论的联系**:佩尔方程的解群 $\{(x, y) : x^2 - Dy^2 = 1\}$ 同构于实二次域 $\mathbb{Q}(\sqrt{D})$ 中整数环的单位群,基本解对应**基本单位元**(Fundamental Unit)。
|
||||
|
||||
4. **广义佩尔方程**:$x^2 - Dy^2 = N$($N \neq \pm 1$)的求解可通过先求 $x^2 - Dy^2 = 1$ 的基本解,再结合特解得到所有解。
|
||||
|
||||
5. **递推时的细节**:需要注意的一点是,n为偶数时,基本解为 $p_{2n-1}, q_{2n-1}$ ,计算时需要重复a序列,需要重复的是 $a_1$ 开始的a序列,那么 $2n-1$ 项就只需要重复到 $2 a_0$ 的前一项。
|
||||
92
solutions/0066.DiophantineEq/euler_66.py
Normal file
92
solutions/0066.DiophantineEq/euler_66.py
Normal file
@@ -0,0 +1,92 @@
|
||||
"""
|
||||
Consider quadratic Diophantine equations of the form:
|
||||
x^2 - D y^2 = 1
|
||||
For example, when D = 13 , the minimal solution in x is $649^2 - 13 * 180^2 = 1$ .
|
||||
It can be assumed that there are no solutions in positive integers when D is square.
|
||||
By finding minimal solutions in x for D={2,3,5,6,7} , we obtain the following:
|
||||
3^2 - 2 * 2^2 = 1
|
||||
2^2 - 3 * 2^2 = 1
|
||||
9^2 - 5 * 4^2 = 1
|
||||
5^2 - 6 * 2^2 = 1
|
||||
8^2 - 7 * 3^2 = 1
|
||||
Hence, by considering minimal solutions in x for D <= 7 ,
|
||||
the largest x is obtained when D = 5 .
|
||||
Find the value of D <= 1000 in minimal solutions of x for
|
||||
which the largest value of x is obtained.
|
||||
"""
|
||||
|
||||
import time
|
||||
from functools import wraps
|
||||
from math import isqrt
|
||||
from typing import Any, Callable, TypeVar
|
||||
|
||||
F = TypeVar("F", bound=Callable[..., Any])
|
||||
|
||||
|
||||
def benchmark(repeat: int = 1) -> Callable[[F], F]:
|
||||
if repeat < 1:
|
||||
raise ValueError("repeat >= 1")
|
||||
|
||||
def decorator(func: F) -> F:
|
||||
@wraps(func)
|
||||
def wrapper(*args: Any, **kwargs: Any) -> Any:
|
||||
total = 0.0
|
||||
result = None
|
||||
|
||||
for _ in range(repeat):
|
||||
start = time.perf_counter()
|
||||
result = func(*args, **kwargs)
|
||||
end = time.perf_counter()
|
||||
total += end - start
|
||||
|
||||
wrapper.avg_time = total / repeat # type: ignore[attr-defined]
|
||||
wrapper.total_time = total # type: ignore[attr-defined]
|
||||
|
||||
print(
|
||||
f"[Benchmark] {func.__name__} | repeated {repeat} times | "
|
||||
f"average: {wrapper.avg_time:.6f}s | total: {wrapper.total_time:.6f}s" # type: ignore[attr-defined]
|
||||
)
|
||||
return result
|
||||
|
||||
wrapper.avg_time = 0.0 # type: ignore[attr-defined]
|
||||
wrapper.total_time = 0.0 # type: ignore[attr-defined]
|
||||
return wrapper # type: ignore[return-value]
|
||||
|
||||
return decorator
|
||||
|
||||
|
||||
def diophantine_eq(D: int) -> int | None:
|
||||
a = [int(isqrt(D))]
|
||||
if a[0] * a[0] == D:
|
||||
return None
|
||||
p = [0, 1, int(isqrt(D))]
|
||||
m = 0
|
||||
d = 1
|
||||
while True:
|
||||
m = d * a[-1] - m
|
||||
d = (D - m * m) // d
|
||||
a.append((a[0] + m) // d)
|
||||
p.append(p[-1] * a[-1] + p[-2])
|
||||
if a[-1] == 2 * a[0]:
|
||||
if len(a) % 2 == 1:
|
||||
return p[-2]
|
||||
else:
|
||||
for ai in a[1:-1]:
|
||||
p.append(p[-1] * ai + p[-2])
|
||||
return p[-1]
|
||||
|
||||
|
||||
@benchmark(repeat=10)
|
||||
def find_minimal_solution(limit: int) -> int:
|
||||
min_x: dict[int, int] = {}
|
||||
for D in range(2, limit + 1):
|
||||
x = diophantine_eq(D)
|
||||
if x is not None:
|
||||
min_x[D] = x
|
||||
return max(min_x, key=min_x.__getitem__)
|
||||
|
||||
|
||||
if __name__ == "__main__":
|
||||
max_D = 1000
|
||||
dd = find_minimal_solution(max_D)
|
||||
print(f"D <= {max_D} |-> D of max x = {dd}")
|
||||
Reference in New Issue
Block a user