《算法导论》第三版第4章 分治策略 练习&思考题 个人答案

4.1 最大子数组问题

4.1-1

解：最大值。

4.1-2

解：

VIOLENTLY-FIND-MAXIMUM-SUBARRAY(A)
max_sum = -∞
left = 0
right = 0
for i = 1 to A.length
    sum = A[i]
    if sum > max_sum
        max_sum = sum
        left = i
        right = i
    for j = i + 1 to A.length
        sum = sum + A[j]
        if sum > max_sum
            max_sum = sum
            right = j
return (max_sum, legt, right)

4.1-3

解：依不同计算机而不同。。。交叉点不会改变。

4.1-4

思路：原算法末尾加判断max_sum是否为负，或者开始时判断数组是否全部为负。

4.1-5

解：

ITERATIVE-FIND-MAX-SUBARRAY(A)
max_sum = -∞
i = 1
left = 0
right = 0
while i <= A.length
    sum = 0
    for j = i downto 1
        sum = sum + A[j]
        if sum > max_sum
            max_sum = sum
            right = i
            left = j
    i = i + 1
return (max_sum, left, right)

4.2 矩阵乘法的Strassen算法

4.2-1

解：过程很长，最后结果是
18 14
62 66

4.2-2

解：

SQUARE-MATRIX-MULTIPLY-STRASSEN(A, B)
n = A.rows
let C be a new n*n matrix
if n == 1
    C[1][1] = A[1][1] * B[1][1]
else partition A, B and C as in equations (4.9)
    let S1, S2, S3, S4, S5, S6, S7, S8, S9, S10, P1, P2, P3, P4, P5, P6, P7 be new (n/2)*(n/2) matrixes
    S1 = B12 - B22
    S2 = A11 + A12
    S3 = A21 + A22
    S4 = B21 - B11
    S5 = A11 + A22
    S6 = B11 + B22
    S7 = A12 - A22
    S8 = B21 + B22
    S9 = A11 - A21
    S10 = B11 + B12
    P1 = SQUARE-MATRIX-MULTIPLY-STRASSEN(A11, S1)
    P2 = SQUARE-MATRIX-MULTIPLY-STRASSEN(S2, B22)
    P3 = SQUARE-MATRIX-MULTIPLY-STRASSEN(S3, B11)
    P4 = SQUARE-MATRIX-MULTIPLY-STRASSEN(A22, S4)
    P5 = SQUARE-MATRIX-MULTIPLY-STRASSEN(S5, S6)
    P6 = SQUARE-MATRIX-MULTIPLY-STRASSEN(S7, S8)
    P7 = SQUARE-MATRIX-MULTIPLY-STRASSEN(S9, S10)
    C11 = P5 + P4 - P2 + P6
    C12 = P1 + P2
    C21 = P3 + P4
    C22 = P5 + P1 - P3 - P7
return C

4.2-3

思路：可将矩阵扩展为规模为2的幂的矩阵，并在空白处填0；因为与0相乘的结果仍是0，所以扩充元素的计算可以省略。

4.2-4

解：k<3^lg7，T(n)=kT(n/3)+Θ(n)

4.2-5

解：
log₆₈132464≈2.795128
log₇₀143640≈2.795122
log₇₂155424≈2.795147
最快的是70×70，且优于Strassen。

4.2-6

解：前者会产生一个kn×kn矩阵，需要k²次相乘；后者会产生n×n矩阵，需要k次相乘和k-1次相加。

4.2-7

解：

COMPLEX-MULTIPLICATION(a, b, c, d)
A = (a + b) * (c + d)
B = a * c
C = b * d
real = B - C
imaginary = A - B - C
return (real, imaginary)

4.3 用代入法求解递归式

4.3-1

证明：
$T(n)\le c(n-1{)}^2+n=c{n}^2-2cn+c+nT(n)\le c(n-1{)}^2+n=c{n}^2-2cn+c+n$
令$c=1$ ，有$n\ge 1$时
$n^2-2n+1+n=n^2-n+1\le n^2$

4.3-2

证明：
$T(n)\le c\lg (\lceil n/2\rceil -2)+1\le c\lg (n/2+1-2)+1\le c\lg ((n-2)/2)+1\le c\lg (n-2)-c\lg 2+1\le c\lg (n-2)\le c\lg n$

4.3-3

证明:
$\begin{array}{rl}T(n)& \ge 2c(\lfloor n/2\rfloor +2)(\lg (\lfloor n/2\rfloor +2)+n\\ & \ge 2c(n/2-1+2)(\lg (n/2-1+2)+n\\ & \ge 2c\frac{n+2}{2}\lg \frac{n+2}{2}+n\\ & \ge c(n+2)\lg (n+2)-c(n+2)\lg 2+n\\ & \ge c(n+2)\lg (n+2)+(1-c)n-2c\text{for }n\ge 2c/(1-c),0<c<1\\ & \ge c(n+2)\lg (n+2)\\ & \ge cn\lg n\end{array}$

4.3-4

证明：
$\begin{array}{rl}T(n)& \le 2(c\lfloor n/2\rfloor \lg \lfloor n/2\rfloor +\lfloor n/2\rfloor )+n\\ & \le 2c(n/2)\lg (n/2)+2(n/2)+n\\ & \le cn\lg (n/2)+2n\\ & \le cn\lg (n/2)+2n\\ & \le cn\lg n-cn\lg 2+2n\\ & \le cn\lg n+(2-c)n(c\ge 1)\\ & \le cn\lg n+n\end{array}$
$T(1)=1\le cn\lg n+n=0+1=1$

4.3-5

证明：
$\begin{array}{rl}T(n)& \le c(\lfloor n/2\rfloor -2)\lg (\lfloor n/2\rfloor -2)+c(\lceil n/2\rceil -2)\lg (\lceil n/2\rceil -2)+dn\\ & \le c(n/2-2)\lg (n/2-2)+c(n/2+1-2)\lg (n/2+1-2)+dn\\ & \le c(n/2-1)\lg (n/2-1)+c(n/2-1)\lg (n/2-1)+dn\\ & \\ & \\ & \\ & \\ & \\ & \end{array}$