Chapter 3 Summations Arithematic Series n 1 2 3 n k k 1 1 n(n 1) (n 2 ) 2 e.g. Insertion Sort Geometric Series (Exponential Series) n 1 x x 2 x3 x n x k k 0 x n1 1 x 1 x 1, n x k k 0 1 1 x n Q. (2k 1) =? k 1 Harmonic Series 1 1 1 1 Hn 2 3 4 n 1 1 1 1 1 1 1 1 k , n 2k 2 2 4 4 4 4 2 1111 (k個) 1 log n O(1) Q. Prove the following: n 1 2k 1 ln( n ) o(1) k 1 Some series can be obtained by integrating or differentiating other series. 1 x kxk x kxk 1 x ( x k ) x( x k ) x( 1) 1 x (1 x) 2 k 0 k 1 k 1 k 1 Telescoping Series n a k 1 k a k 1 a n a 0 e.g. n 1 n 1 1 1 1 1 k (k 1) k k 1 1 n k 1 k 1 Products n a k k 1 n n k 1 k 1 lg( a k ) lg a k Problem: n S 2 4k k 1 n n n n k 1 k 1 k 1 k 1 lg S lg( 2 4 k ) lg 2 lg 4 k lg 2 2k n lg 2 (1 n) n n(lg 2 1 n) n (2 n) Q. Show a close form of the following product. n 1 (1 k k 2 2 ) Bounding Summations 1. Mathematical Induction Problem. To prove n 1 k 2 n(n 1) k 1 It holds for n=1 Suppose it holds for n, we will show that it holds for n+1 n 1 n k 1 k 1 1 1 k k (n 1) 2 n(n 1) (n 1) 2 (n 1)(n 2) n 3 k Problem: Prove c 3n k 0 It holds for n=1 Suppose it holds for n, we will show that it holds for n+1 n 1 3 k k 0 c 1 1 c 3n 3n 1 ( 1) 3n 1 c 3n 1 ( ) c 3n 1 3 3 c as long as 1 1 3 1 i.e. c 3 c 2 A fallacious proof for n k O ( n) k 1 It holds for n=1. Suppose it holds for n, we show that it holds for n+1 as follows: n 1 n k 1 k 1 k k (n 1) O(n) n 1 O(n 1) This is incorrect because n 1 n k 1 k 1 k k n 1 c n (n 1) c (n 1) c (n 1) NOT c (n 1) 2. Bounding the terms n n k n n k 1 k 1 n a k 1 k when n a k 0 k n amax a k 1 r 1 ak n a0 n a a k 0 k k 0 0 r k a 0 r k a0 k 0 1 1 r Problem: Prove k 3 k 1 2 k k 1 a k 1 3 k 1 k 1 1 2 k ak k 3 3 3k for k 1 k 3 Thus k 1 1 1 1 3 1 2 3 3 1 3 Note that this e.g. k ak 1 must be a constant, rather than a bound. ak 1 k k 1 a k 1 k 1 ak k 1 But you cannot conclude that this summation converges. 3. Splitting summations n k 0 1 k 0 k 0 ak ak n ak (1) k k0 n a k k0 k e.g. a k 1 ak 2 k2 k2 k2 k O(1) k k k 0 2 k 0 2 k 3 2 (k 1) 2 k 1 (k 1) 2 8 22 9 k 2k2 k 2 when k 3 lg n 2 1 1 lg n 1 lg n 2 1 1 i 1 lg n 1 e.g. i k 1 k i 0 j 0 2 j i 0 j 0 2 i 0 i n i lg n Q. show the asymptotic upper bound is n k k 0 2 (by using bounding each term and letting n+R be a power of 2) 4. Approximation by Integral See Fig. 3.1 n n 1 m 1 m n f ( x)dx f (k ) n 1 k m m n n f ( x)dx f (k ) k m m 1 f ( x)dx if f(x) is a monotonically increasing function. f ( x)dx if f(x) is a monotonically decreasing function. n 1 1 n 1 1 dx k k 1 k n 1 k 1 k ln( n 1) n 1 n 1 k 1 k 1 k 1 n 1 k 2 1 dx 1 ln n k n Q. Approximate k k 1 Exercise: 3.2-1. 3 w/ an integral Chapter 4 Recurrences Recall the execution T(n) for Merge Sort (1) T ( n) n 2T ( 2 ) (n) if n 1 n 1 For now, ignore the floor and ceiling. Substitution Method 1. Make a good guess 2. Substitute the values in the recurrence by using the guess 3. Verify its validity on both the recurrence and the boundary. Problem: n T (n) 2T ( ) n 2 Guess T (n) c n lg n n n T (n) 2 c lg n c n(lg n lg 2) n c n lg n c n n c n lg n 2 2 T (1) 1 T (1) c 1 lg 1 0 NOT VALID! Use T(2) as the boundary condition T (2) 2 2 4 T (2) c 2 lg 2 2 c when c 2 How to make a good guess? Problem: n T (n) 2T ( 17) n 2 We can guess T (n) c n lg n Exercise. Prove the above by using substitution method. Add terms of lower order Problem: c 1 n n T ( n) T ( ) T ( ) 1 2 2 It is natural to guess T (n) cn. n n T (n) c c 1 c n 1 NOT cn 2 2 Guess T (n) cn b n n T (n) (c b) (c b) 1 cn 2b 1 cn b 2 2 b 1 Pitfalls Guess T (n) O (n) n T (n) 2(c ) n cn n O(n) 2 Wrong Changing Variables T (n) 2T n lg n Let m lg n m 2 T (2 ) 2T (2 ) m m m S ( m) 2 S ( ) m 2 S (m) O(m lg m) T (n) O(lg n lg lg n) Exercise: 4.1-2. Iteration Method n n n n n n T (n) 3T ( ) n n 3( 3T ) n 3 9( 3T ) 4 4 16 4 16 64 3 9 27 n n n n 3 9 27 n n n n 3log4 n 4 16 64 4 16 64 3 n ( )i (n log4 3 ) 4n O(n) O(n) i 0 4 Recursive Trees n T (n) 2T ( ) n 2 see Fig. 4.1 2 n 2n T ( n) T ( ) T ( ) n 3 3 see Fig. 4.2 Q: Use iteration to solve T (n) T (n a) T (a) n n Exercise: Use recursion tree to solve T (n) 4T ( ) n 2 Generating Functions (1 ax)(1 bx)(1 cx) 1 (a b c) x (ab bc ac) x 2 abcx 3 n n n n n (1 x) n ( ) ( ) x ( ) x 2 ( ) x 3 ( ) x n 0 1 2 3 n n n n n e.g. ( ) ( ) ( ) ( ) 2 n 0 1 2 n Problem: n n n ( ) 2 ( ) 2 ( ) 2 =? 0 1 n Method 1: It is the constant term of 1 (1 x) n (1 ) n x n (1 x) 2 n x 2n coefficient of x n ( ) n Method 2. Use constructive combinatorics to choose n objects from 2n objects choose i choose n-i n obj Q. Prove that n n n n ( ) 2( ) 3( ) n( ) n 2 n 1 1 2 3 n n n n n ( ) ( ) x ( ) x 2 ( ) x n (1 x) n 0 1 2 n differentiating both sides Characteristic Equations Linear Recurrence Relation n obj an an1 an2 (Fibonacci Seq) Its characteristic equation is 2 1 0 where is the characteristic root. 1 5 1 5 , 2 2 2 The total solution is 1 1 5 n 1 5 n ) A2 ( ) 2 2 a 0 0 A1 A2 a n A1 ( 1 5 1 5 A2 2 2 1 1 A1 , A2 5 5 a1 1 A1 Problem: an 6an 1 12an 2 8an 3 0 a0 1, a1 2, a2 8 3 6 2 12 8 0 ( 2) 3 0 an ( A1n 2 A2 n A3 )( 2) n 1 1 A1 , A2 , A3 1 2 2 Problem: an 2an1 n 3 , a0=0. Solve an 2an1 0 gives the homogenous solution For the particular solution Assume an B n D 3B n 3D 2 B n 3 1 11 B , D 3 9 n 11 Thus an A (2) n 3 9 16 a0 3 A 9 Generating Function A(2) n Let A(x) be the generating function of the sequence (a0 , a1 , a2 ,) i.e. A( x) a 0 a1 x a 2 x 2 a n x n Problem: Solve an an1 2(n 1) , a0=2. n 1 n 1 n 1 an x n an1 x n 2 (n 1) x n A( x) a 0 x A( x) A( x) 2x 2 (1 x) 2 a0 2 2x 2 2 3 (1 x) 1 x n2 1 (1 x) 3 term is (n 2)! 1 之 x n2 3 (1 x) Therefore a n n(n 1) 2 3 4 5 n n(n 1) (n 2)! 2 Q. Using generating function to solve an 2an1 n 3 Problem: Solve the following recurrence equations: an ,r an 1,r 1 an 1,r an , 0 1 n 0, an ,r 0 n r Fn ( x) an , 0 an ,1 x an , 2 x 2 an ,r x r r 1 r 1 r 1 an,r x r an1,r 1 x r an1,r x r Fn ( x) an , 0 xFn 1 ( x) Fn 1 ( x) an 1, 0 Fn ( x) (1 x) Fn 1 ( x) Fn ( x) (1 x) Fn 1 ( x) (1 x) n a0, 0 (1 x) n n an , r ( ) r Exercise: Solve a n ,r a n ,r 1 a n 1,r a n , 0 1, a n ,r 0 if n r