多项式求值秦九韶算法

1.多项式求值的秦九韶算法

设给定$n$次多项式$$p(x)=a_0{x}^n+a_1{x}^{n-1}+\cdots +a_{n-1}x+a_n,a_0\ne 0,$$
求$x^{\ast }处的值p(x^{\ast }).若直接计算每一项a_i{x}^{n-i}再相加,共需要$
$$\sum _{i=0}^n(n-i)=1+2+\cdots +n=\frac{n(n+1)}{2}=O(n^2)$$次乘法，$n$次加法.若采用
$$p(x)=(\cdots (a_{0}x+a_1)x+\cdots +a_{n-1})x+a_n,$$它可表示为
$$\{\begin{array}{l}{b}_0=a_0,\\ b_i=b_{i-1}{x}^{\ast }+a_i,i=1,2,\dots ,n,\end{array}$$
则$b_n=p(x^{\ast })即为所求.$此算法称为秦九韶算法，用它计算$n次多项式p(x)$的值只用$n次乘法和n次加法$
$乘法次数由O(n^2)降为O(n),且只用n+2个存储单元，国外称此算法为Horner算法$

2.Python实现秦九韶算法

import numpy as np
from sympy.abc import x


def polynomial(coefficient_array: np.ndarray, is_ascending_order: bool = False):
    """
    构造指定系数的多项式
    :param coefficient_array: 多项式系数数组
    :param is_ascending_order: 多项式排列顺序
    :return: 多项式
    """
    degree = coefficient_array.shape[0]
    if is_ascending_order:
        coefficient_array = coefficient_array[::-1]
    # default,descending order arrangement
    p_x = coefficient_array[-1]
    for i in range(degree - 1, 0, -1):
        p_x += coefficient_array[degree - i - 1] * pow(x, i)
    return p_x


def qin_jiu_shao(coefficient_array: np.ndarray, value, is_ascending_order: bool = False):
    """
    多项式求值的秦九韶算法,即Horner算法
    :param coefficient_array: 多项式系数数组
    :param value: 自变量的值
    :param is_ascending_order: 多项式排列顺序
    :return: the value of polynomial
    example:
    2*x**4 - 3*x**2 + 3*x - 4=((((2x+0)x-3)x+3)x-4)
    3*x**5 - 2*x**3 + x + 7=(((((3x+0)x-2)x+0)x+1)x+7)
    """
    if is_ascending_order:
        coefficient_array = coefficient_array[::-1]
    degree = coefficient_array.shape[0]
    b = coefficient_array[0]
    for i in range(1, degree):
        b = b * value + coefficient_array[i]
    return b

3.多项式求值测试

if __name__ == '__main__':
    # 测试成功，来源详见李庆扬数值分析第5版P14，e.g.11
    # 降幂排列
    coefficient0 = np.array([2, 0, -3, 3, -4])
    print(polynomial(coefficient0))
    print(qin_jiu_shao(coefficient0, -2))
    # 升幂排列
    coefficient1 = np.array([-4, 3, -3, 0, 2])
    print(polynomial(coefficient1, is_ascending_order=True))
    print(qin_jiu_shao(coefficient1, -2, is_ascending_order=True))
    # 测试成功，来源详见李庆扬数值分析第5版P20，e.x.14
    # 降幂排列
    coefficient2 = np.array([3, 0, -2, 0, 1, 7])
    print(polynomial(coefficient2))
    print(qin_jiu_shao(coefficient2, 3))
    # 升幂排列
    coefficient3 = np.array([7, 1, 0, -2, 0, 3])
    print(polynomial(coefficient3, is_ascending_order=True))
    print(qin_jiu_shao(coefficient3, 3, is_ascending_order=True))

4.求值运行结果