Approximate Counting via the Poisson-Laplace-Mellin Method (joint with Chung-Kuei Lee and Helmut Prodinger)
|
|
- Basil Lawson
- 5 years ago
- Views:
Transcription
1 Approximate Counting via the Poisson-Laplace-Mellin Method (joint with Chung-Kuei Lee and Helmut Prodinger) Michael Fuchs Department of Applied Mathematics National Chiao Tung University Hsinchu, Taiwan Montreal, June 18, 2012 Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
2 Approximate Counting (Morris 1978) Space needed for counting n objects: Θ(log n). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
3 Approximate Counting (Morris 1978) Space needed for counting n objects: Θ(log n). Problem: What if space is very limited? Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
4 Approximate Counting (Morris 1978) Space needed for counting n objects: Θ(log n). Problem: What if space is very limited? Answer: Allow an error tolerance: approximate counting. Counter C n with C 0 = 0 and (0 < q < 1) { C n + 1, with probability q Cn ; C n+1 = C n, with probability 1 q Cn. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
5 Approximate Counting (Morris 1978) Space needed for counting n objects: Θ(log n). Problem: What if space is very limited? Answer: Allow an error tolerance: approximate counting. Counter C n with C 0 = 0 and (0 < q < 1) { C n + 1, with probability q Cn ; C n+1 = C n, with probability 1 q Cn. Easy to show: E(q Cn ) = n ( q 1 1 ) + 1. Now, only Θ(log log n) space is needed. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
6 Applications Approximate counting has found many applications: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
7 Applications Approximate counting has found many applications: Analysis of the Webgraph. Monitoring network traffic. Finding patterns in protein and DNA sequencing. Computing frequency moments of data streams. Data storage in flash memory. Etc. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
8 Applications Approximate counting has found many applications: Analysis of the Webgraph. Monitoring network traffic. Finding patterns in protein and DNA sequencing. Computing frequency moments of data streams. Data storage in flash memory. Etc. Many refinements have been proposed. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
9 The Markov Chain C n Other problems leading to the same Markov chain: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
10 The Markov Chain C n Other problems leading to the same Markov chain: Width of greedy decomposition of random acyclic digraphs into node-disjoint paths. Size of greedy independent set in random graphs. Size of greedy clique in random graphs. Length of leftmost path in random digital search trees. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
11 The Markov Chain C n Other problems leading to the same Markov chain: Width of greedy decomposition of random acyclic digraphs into node-disjoint paths. Size of greedy independent set in random graphs. Size of greedy clique in random graphs. Length of leftmost path in random digital search trees. Variations of this Markov chain were also studied: Simon (1988); Crippa and Simon (1997); Bertoin, Biane and Yor (2003); Guillemin, Robert and Zwart (2004); Louchard and Prodinger (2008) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
12 Analysis of Approximate Counting Flajolet (1985): E(C n ) log 1/q n + C mean + F (log 1/q n), where F (z) is a 1-periodic function Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
13 Analysis of Approximate Counting Flajolet (1985): E(C n ) log 1/q n + C mean + F (log 1/q n), where F (z) is a 1-periodic function and Var(C n ) C var + G(log 1/q n), where G(z) is a 1-periodic function and C var = π 2 6 log 2 (1/q) α β log(1/q) l sinh(2lπ 2 / log(1/q)) l 1 with α = l 1 ql /(1 q l ) and β = l 1 q2l /(1 q l ) 2. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
14 Methods Many different methods have been used: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
15 Methods Many different methods have been used: Mellin Transform: Flajolet (1985); Prodinger (1992) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
16 Methods Many different methods have been used: Mellin Transform: Flajolet (1985); Prodinger (1992) Rice Method: Kirschenhofer and Prodinger (1991) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
17 Methods Many different methods have been used: Mellin Transform: Flajolet (1985); Prodinger (1992) Rice Method: Kirschenhofer and Prodinger (1991) Euler Transform: Prodinger (1994) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
18 Methods Many different methods have been used: Mellin Transform: Flajolet (1985); Prodinger (1992) Rice Method: Kirschenhofer and Prodinger (1991) Euler Transform: Prodinger (1994) Analysis of Extreme Value Distributions: Louchard and Prodinger (2006) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
19 Methods Many different methods have been used: Mellin Transform: Flajolet (1985); Prodinger (1992) Rice Method: Kirschenhofer and Prodinger (1991) Euler Transform: Prodinger (1994) Analysis of Extreme Value Distributions: Louchard and Prodinger (2006) Martingale Theory: Rosenkrantz (1987) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
20 Methods Many different methods have been used: Mellin Transform: Flajolet (1985); Prodinger (1992) Rice Method: Kirschenhofer and Prodinger (1991) Euler Transform: Prodinger (1994) Analysis of Extreme Value Distributions: Louchard and Prodinger (2006) Martingale Theory: Rosenkrantz (1987) Probability Theory: Robert (2005) Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
21 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
22 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
23 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
24 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
25 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
26 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
27 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
28 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
29 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
30 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
31 Digital Search Tree (DST) Introduced by Coffman and Eve (1970). Example: a digital search tree build from 9 keys: Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
32 Random Model and Leftmost Path Random Model: Bits are generated by independent Bernoulli random variables with mean p. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
33 Random Model and Leftmost Path Random Model: Bits are generated by independent Bernoulli random variables with mean p. Two types of trees: p = 1/2: symmetric digital search tree; Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
34 Random Model and Leftmost Path Random Model: Bits are generated by independent Bernoulli random variables with mean p. Two types of trees: p = 1/2: symmetric digital search tree; p 1/2: asymmetric digital search tree. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
35 Random Model and Leftmost Path Random Model: Bits are generated by independent Bernoulli random variables with mean p. Two types of trees: p = 1/2: symmetric digital search tree; p 1/2: asymmetric digital search tree. Length of the Leftmost Path: X n : number of vertices on leftmost path. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
36 Random Model and Leftmost Path Random Model: Bits are generated by independent Bernoulli random variables with mean p. Two types of trees: p = 1/2: symmetric digital search tree; p 1/2: asymmetric digital search tree. Length of the Leftmost Path: X n : number of vertices on leftmost path. Note that: X n d = Cn. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
37 Distributional Recurrence of X n X n+1 d = XIn + 1 Root 0 1 I n d = Binomial(n, q); X n, I n independent. Size: I n Size: n I n Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
38 Distributional Recurrence of X n X n+1 d = XIn + 1 Root 0 1 I n d = Binomial(n, q); X n, I n independent. Size: I n Size: n I n Recurrence of moments: f n+1 = n j=0 ( ) n q j p n j f j + g n. j Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
39 Other Shape Parameters Depth Konheim, Newman, Knuth, Devroye, Louchard, Szpankowski Total Path Length Flajolet, Sedgewick, Prodinger, Kirschenhofer, Szpankowski, Hubalek Peripheral Path Length Drmota, Gittenberger, Panholzer, Prodinger, Ward # of Occurrences of Patterns Knuth, Flajolet, Sedgewick, Prodinger, Kirschenhofer Colless Index Fuchs, Hwang, Zacharovas Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
40 Methods Rice Method Introduced by Flajolet and Sedgewick. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
41 Methods Rice Method Introduced by Flajolet and Sedgewick. Approach of Flajolet and Richmond Based on Euler transform, Mellin transform, and singularity analysis. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
42 Methods Rice Method Introduced by Flajolet and Sedgewick. Approach of Flajolet and Richmond Based on Euler transform, Mellin transform, and singularity analysis. Approach via Analytic Depoissonization Introduced by Jacquet & Regnier and Jacquet & Szpankowski. Based on saddle point method and Mellin transform. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
43 Methods Rice Method Introduced by Flajolet and Sedgewick. Approach of Flajolet and Richmond Based on Euler transform, Mellin transform, and singularity analysis. Approach via Analytic Depoissonization Introduced by Jacquet & Regnier and Jacquet & Szpankowski. Based on saddle point method and Mellin transform. Schachinger s Approach Largely elementary. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
44 Poisson-Laplace-Mellin Method Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
45 Poisson-Laplace-Mellin Method Poissonized mean and variance satisfy differential-function equations. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
46 Poisson-Laplace-Mellin Method Poissonized mean and variance satisfy differential-function equations. Due to Jacquet-Szpankowski s theory of depoissonization only the Poisson model has to be analyzed. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
47 Poisson-Laplace-Mellin Method Poissonized mean and variance satisfy differential-function equations. Due to Jacquet-Szpankowski s theory of depoissonization only the Poisson model has to be analyzed. Laplace transform to get rid of the differential operator. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
48 Poisson-Laplace-Mellin Method Poissonized mean and variance satisfy differential-function equations. Due to Jacquet-Szpankowski s theory of depoissonization only the Poisson model has to be analyzed. Laplace transform to get rid of the differential operator. A normalization factor simplifies the functional equation satisfied by Laplace transform. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
49 Poisson-Laplace-Mellin Method Poissonized mean and variance satisfy differential-function equations. Due to Jacquet-Szpankowski s theory of depoissonization only the Poisson model has to be analyzed. Laplace transform to get rid of the differential operator. A normalization factor simplifies the functional equation satisfied by Laplace transform. Mellin transform is applied which can be computed explicitly. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
50 Poisson-Laplace-Mellin Method Poissonized mean and variance satisfy differential-function equations. Due to Jacquet-Szpankowski s theory of depoissonization only the Poisson model has to be analyzed. Laplace transform to get rid of the differential operator. A normalization factor simplifies the functional equation satisfied by Laplace transform. Mellin transform is applied which can be computed explicitly. We use inverse Mellin transform and inverse Laplace transform to obtain asymptotic expansions in the Poisson model. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
51 Poissonization Moments satisfy the recurrence: f n+1 = n j=0 ( ) n q j p n j f j + g n. j Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
52 Poissonization Moments satisfy the recurrence: f n+1 = n j=0 ( ) n q j p n j f j + g n. j Consider Poisson-generating function of f n and g n, i.e., f(z) := e z n 0 z n f n n!, g(z) := z n e z g n n!. n 0 Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
53 Poissonization Moments satisfy the recurrence: f n+1 = n j=0 ( ) n q j p n j f j + g n. j Consider Poisson-generating function of f n and g n, i.e., Then, f(z) := e z n 0 z n f n n!, g(z) := z n e z g n n!. n 0 f(z) + f (z) = f(qz) + g(z). This is a differential-functional equation. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
54 Poisson Heuristic Poisson Heuristic: f n sufficiently smooth = f n f(n). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
55 Poisson Heuristic Poisson Heuristic: f n sufficiently smooth = f n f(n). More precisely: if f n is smooth enough, f n j 0 f (j) (n) τ j (n) = n! f(n) n 2 f (n) +..., where τ j (n) := n![z n ](z n) j e z This is called Poisson-Charlier expansion (can be already found in Ramanujan s notebooks). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
56 Jacquet-Szpankowski-admissibility (JS-admissibility) f(z) is called JS-admissible if (I) Uniformly for arg(z) ɛ, f(z) = O (O) Uniformly for ɛ < arg(z) π, ( ) z α log β z, f(z) := e z f(z) = O (e (1 ɛ) z ). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
57 Jacquet-Szpankowski-admissibility (JS-admissibility) f(z) is called JS-admissible if (I) Uniformly for arg(z) ɛ, f(z) = O (O) Uniformly for ɛ < arg(z) π, Theorem (Jacquet and Szpankowski) If f(z) is JS-admissible, then ( ) z α log β z, f(z) := e z f(z) = O (e (1 ɛ) z ). f n f(n) n 2 f (n) +. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
58 Depoissonization JS-admissibility satisfies closure properties: (i) f, g JS-admissible, then f + g JS-admissible. (ii) f JS-admissible, then f JS-admissible. Etc. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
59 Depoissonization JS-admissibility satisfies closure properties: (i) f, g JS-admissible, then f + g JS-admissible. (ii) f JS-admissible, then f JS-admissible. Etc. Proposition Consider f(z) + f (z) = f(qz) + g(z). We have, g(z) JS-admissible f(z) JS-admissible. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
60 Poissonized Mean and Second Moment Define f 1 (z) = e z n 0 E(X n ) zn n!, f2 (z)e z E(Xn) 2 zn n! n 0 which are poissonized mean and second moment. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
61 Poissonized Mean and Second Moment Define f 1 (z) = e z n 0 E(X n ) zn n!, f2 (z)e z E(Xn) 2 zn n! n 0 which are poissonized mean and second moment. Then, f 1 (z) + f 1(z) = f 1 (qz) + 1 f 2 (z) + f 2(z) = f 2 (qz) + 2 f 1 (qz) + 1 Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
62 Poissonized Mean and Second Moment Define f 1 (z) = e z n 0 E(X n ) zn n!, f2 (z)e z E(Xn) 2 zn n! n 0 which are poissonized mean and second moment. Then, f 1 (z) + f 1(z) = f 1 (qz) + 1 f 2 (z) + f 2(z) = f 2 (qz) + 2 f 1 (qz) + 1 Previous results show that f 1 (z), f 2 (z) are JS admissible. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
63 Poissonized Variance Important in order to avoid having to cope with cancelations! Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
64 Poissonized Variance Important in order to avoid having to cope with cancelations! Mean, Variance = Θ(log β n): Ṽ (z) = f 2 (z) f 1 (z) 2. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
65 Poissonized Variance Important in order to avoid having to cope with cancelations! Mean, Variance = Θ(log β n): Ṽ (z) = f 2 (z) f 1 (z) 2. Mean, Variance = Θ(n log β n): Ṽ (z) = f 2 (z) f 1 (z) 2 z f 1(z) 2. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
66 Poissonized Variance Important in order to avoid having to cope with cancelations! Mean, Variance = Θ(log β n): Ṽ (z) = f 2 (z) f 1 (z) 2. Mean, Variance = Θ(n log β n): Ṽ (z) = f 2 (z) f 1 (z) 2 z f 1(z) 2. More general: Ṽ (z) = f 2 (z) n 0 f (n) 1 (z) 2 zn n!. Then one obtains even identities! Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
67 Laplace and Mellin Transform (i) We start from, f(z) + f (z) = f(qz) + g(z). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
68 Laplace and Mellin Transform (i) We start from, f(z) + f (z) = f(qz) + g(z). Applying Laplace transform, (s + 1)L [ f(z); s] = 1 [ q L f(z); 1 ] + L [ g(z); s]. q Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
69 Laplace and Mellin Transform (i) We start from, f(z) + f (z) = f(qz) + g(z). Applying Laplace transform, (s + 1)L [ f(z); s] = 1 [ q L f(z); 1 ] + L [ g(z); s]. q Define, Q(s) := l 1 ( ) 1 q l s and Q := Q(1). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
70 Laplace and Mellin Transform (ii) Set f(s) := L [ f(z); s] L [ g(z); s], ḡ(s) := Q( s) Q( s/q). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
71 Laplace and Mellin Transform (ii) Set Then, f(s) := L [ f(z); s] L [ g(z); s], ḡ(s) := Q( s) Q( s/q). f(s) = 1 q f ( ) s + ḡ(s). q Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
72 Laplace and Mellin Transform (ii) Set Then, f(s) := L [ f(z); s] L [ g(z); s], ḡ(s) := Q( s) Q( s/q). f(s) = 1 q f ( ) s + ḡ(s). q Applying Mellin transform, M [ f(s); ω] = M [ḡ(s); ω] 1 q ω 1. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
73 Laplace and Mellin Transform (ii) Set Then, f(s) := L [ f(z); s] L [ g(z); s], ḡ(s) := Q( s) Q( s/q). f(s) = 1 q f ( ) s + ḡ(s). q Applying Mellin transform, M [ f(s); ω] = M [ḡ(s); ω] 1 q ω 1. From this, an asymptotic expansion of f(z) as z is obtained via inverse Mellin transform and inverse Laplace transform. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
74 Inverse Laplace Transform Theorem (F., Hwang, Zacharovas) Let the Laplace transform of f(z) exist and be analytic in C \ (, 0]. Assume that L [ f; s] = O ( s α) uniformly for s 0 and arg(s) π ɛ. Then, f(z) = O ( z α 1) uniformly for z and arg(z) π/2 ɛ. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
75 Our Approach vs. Flajolet-Richmond EGF f(z) Laplace transform of e z f(z) = Euler transform of A(z) OGF A(z) asymptotics of f(z) as z asymptotics of Laplace Q( s) Laplace Q( s) Euler Q( s) asymptotics of Euler Q( s) asymptotics of A(z) as z 1 de-poi by saddle-point Mellin transform singularity analysis Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
76 Main Result Theorem (F., Lee, Prodinger) We have, Var(C n ) k g k n χ k, where g k = Q LΓ(1 + χ k ) h,l,j 0 ( 1) j q h+l+(j+1 2 ) Q h Q l Q j ϕ(χ k ; q h+j + q l+j ). Here, χ k = 2kπi/L, L = log(1/q), Q j = j l=1 (1 ql ) and { π(x χ 1)/(sin(πχ)(x 1)), if x 1; ϕ(χ; x) = πχ/ sin(πχ), if x = 1. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
77 An Identity Corollary (F., Lee, Prodinger) We have, Q L h,l,j 0 ( 1) j q h+l+(j+1 2 ) Q h Q l Q j ψ(q h+j + q l+j ) = π2 6L 2 α β L l sinh(2lπ 2 /L), l 1 where ψ(x) = { log x/(x 1), if x 1; 1, if x = 1. We have a direct proof for this using tools from q-analysis. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
78 Total Path Length (i) T n : total path length in symmetric digital search tree. Theorem (F., Hwang, Zacharovas) We have Var(T n ) n(c var + G(log 2 n)), where G(z) is a 1-periodic function with zero average value and C var = Q L j,h,l 0 ( 1) j 2 (j+1 2 ) Q j Q h Q l 2 h+l δ(2 j h + 2 j l ), where δ(x) := { (x log x 1)/(x 1) 2, if x 1; 1/2, if x = 1. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
79 Total Path Length (ii) Variance of total path length was also derived by Kirschenhofer, Prodinger and Szpankowski with different expression for C var. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
80 Total Path Length (ii) Variance of total path length was also derived by Kirschenhofer, Prodinger and Szpankowski with different expression for C var. To describe their expression we need: Let [F G] 0 denote the 0-th Fourier coefficient of the product of the two Fourier series F (z) and G(z). Put and F (z) = 1 Γ( 1 χ l )e 2lπiz L H(z) = 1 L l 0 ( l 0 1 χ l 2 ) Γ( χ l )e 2lπiz. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
81 C var = 28 3L l 1 l2 l (2 l 1) L l 1 2 ( 1) l+1 (l 5) L (l + 1)l(l 1)(2 l 1) l 3 2Q(1) L + l l Q l 1 2 l 1 + π2 2L L ( ) ( 1) l 2 l+1 2 L(1 2 l+1 )/2 1 L 1 2 l ( 1) r+1 r(r 1)(2 r+l 1) l 1 r 2 ( ) ( 1) r 2 r+1 2 ( j j+r+l+2 1 Q r+l 2 Q r 0 r ( l 1 ( l + 1 ) 2 l+1 2l i i=2 2 (1 2 l r ) 2 + 2l + 2 (1 2 1 l r ) 2 2 L l+1 ( l + 1 ) 2 j j=2 1 2 r+j L j=0 i r+i l r + 2 l+1 ( l + 1 ) L j j=1 l+1 ( l + 1 ) ( 1) i ) j (i + 1)(2 r+j+i 1) + l 1 ( l + 1 ) Qr 2 Q l r 1 1 r 2 l 3 r=2 l Q l 2 j 1 2[F H] 0 [F 2 ] 0. j l+1 ) 1 2 r+j 1 Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
82 Approximate Counting with m Counters (i) Cichoń and Macyna: Consider m counters. When counting the n-th object choose one uniform at random. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
83 Approximate Counting with m Counters (i) Cichoń and Macyna: Consider m counters. When counting the n-th object choose one uniform at random. D n : sum of all counters after counting n objects. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
84 Approximate Counting with m Counters (i) Cichoń and Macyna: Consider m counters. When counting the n-th object choose one uniform at random. D n : sum of all counters after counting n objects. Then, where D n d = C (1) I C (m) I m, P (I 1 = n 1,..., I m = n m ) = 1 ( ) n m n. n 1,..., n m Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
85 Approximate Counting with m Counters (i) Cichoń and Macyna: Consider m counters. When counting the n-th object choose one uniform at random. D n : sum of all counters after counting n objects. Then, where D n d = C (1) I C (m) I m, P (I 1 = n 1,..., I m = n m ) = 1 ( ) n m n. n 1,..., n m Let f D, f C denote Poisson mean of D n and C n. Similar, let ṼD and ṼC denote Poisson variance of D n and C n. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
86 Approximate Counting with m Counters (ii) Poisson model: f D (z) = m f C (z/m), Ṽ D (z) = mṽc(z/m). Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
87 Approximate Counting with m Counters (ii) Poisson model: f D (z) = m f C (z/m), Ṽ D (z) = mṽc(z/m). From this, we obtain the following result. Theorem (F., Lee, Prodinger) We have, E(D n ) m log 1/q (n/m) + mc mean + mf (log 1/q (n/m)), Var(D n ) mc var + mg(log 1/q (n/m)), where C mean, C var and F (z), G(z) are as before. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
88 Approximate Counting with m Counters (iii) Another variant of approximate counting with m counters: Consider m counters. Use a counter until it is increased; then cyclically move on to the next. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
89 Approximate Counting with m Counters (iii) Another variant of approximate counting with m counters: Consider m counters. Use a counter until it is increased; then cyclically move on to the next. D n : sum of all counters after counting n objects. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
90 Approximate Counting with m Counters (iii) Another variant of approximate counting with m counters: Consider m counters. Use a counter until it is increased; then cyclically move on to the next. D n : sum of all counters after counting n objects. Then, where D n d = Xn, X n+m d = XIn + m. X n is the length of the leftmost path in random bucket digital search tree with bucket size m. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
91 Approximate Counting with m Counters (iv) Recurrence of moments: f n+m = n j=0 ( ) n q j p n j f j + g n. j Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
92 Approximate Counting with m Counters (iv) Recurrence of moments: f n+m = n j=0 ( ) n q j p n j f j + g n. j Poissonized variance Ṽ (z) satisfies the differential function equation: m i=0 ( ) m Ṽ (i) (z) = Ṽ (qz) + g(z), i where g(z) = ( m i=0 ( ) ) 2 m f (i) 1 i (z) m i=0 ( ) m ( f1 (z) 2) (i). i Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
93 Approximate Counting with m Counters (v) Theorem (F., Lee, Prodinger) We have, Var(D n ) k g k n χ k, where g k = 1 s χ ( k LΓ(1 + χ k ) 0 Q( s/q) m p(s) + 0 e sz g(z)dz ) ds and p(s) = (s + 1)m 1 ms s 2. Michael Fuchs (NCTU) Approximate Counting Montreal, June 18, / 32
Asymptotic and Exact Poissonized Variance in the Analysis of Random Digital Trees (joint with Hsien-Kuei Hwang and Vytas Zacharovas)
Asymptotic and Exact Poissonized Variance in the Analysis of Random Digital Trees (joint with Hsien-Kuei Hwang and Vytas Zacharovas) Michael Fuchs Institute of Applied Mathematics National Chiao Tung University
More informationAsymptotic variance of random symmetric digital search trees
Discrete Mathematics and Theoretical Computer Science DMTCS vol. :,, 3 66 Asymptotic variance of random symmetric digital search trees Hsien-Kuei Hwang, Michael Fuchs and Vytas Zacharovas Institute of
More informationDependence between Path Lengths and Size in Random Trees (joint with H.-H. Chern, H.-K. Hwang and R. Neininger)
Dependence between Path Lengths and Size in Random Trees (joint with H.-H. Chern, H.-K. Hwang and R. Neininger) Michael Fuchs Institute of Applied Mathematics National Chiao Tung University Hsinchu, Taiwan
More informationThe variance for partial match retrievals in k-dimensional bucket digital trees
The variance for partial match retrievals in k-dimensional ucket digital trees Michael FUCHS Department of Applied Mathematics National Chiao Tung University January 12, 21 Astract The variance of partial
More informationThe Subtree Size Profile of Plane-oriented Recursive Trees
The Subtree Size Profile of Plane-oriented Recursive Trees Michael Fuchs Department of Applied Mathematics National Chiao Tung University Hsinchu, Taiwan ANALCO11, January 22nd, 2011 Michael Fuchs (NCTU)
More informationROSENA R.X. DU AND HELMUT PRODINGER. Dedicated to Philippe Flajolet ( )
ON PROTECTED NODES IN DIGITAL SEARCH TREES ROSENA R.X. DU AND HELMUT PRODINGER Dedicated to Philippe Flajolet (98 2) Abstract. Recently, 2-protected nodes were studied in the context of ordered trees and
More informationOn the number of subtrees on the fringe of random trees (partly joint with Huilan Chang)
On the number of subtrees on the fringe of random trees (partly joint with Huilan Chang) Michael Fuchs Institute of Applied Mathematics National Chiao Tung University Hsinchu, Taiwan April 16th, 2008 Michael
More informationForty years of tree enumeration
October 24, 206 Canada John Moon (Alberta) Counting labelled trees Canadian mathematical monographs What is our subject? Asymptotic enumeration or Probabilistic combinatorics? What is our subject? Asymptotic
More informationAnalytic Depoissonization and Its Applications
Analytic Depoissonization and Its Applications W. Szpankowski Department of Computer Science Purdue University W. Lafayette, IN 47907 April 29, 2008 AofA and IT logos This research is supported by NSF
More informationA Note on a Problem Posed by D. E. Knuth on a Satisfiability Recurrence
A Note on a Problem Posed by D. E. Knuth on a Satisfiability Recurrence Dedicated to Philippe Flajolet 1948-2011 July 4, 2013 Philippe Jacquet Charles Knessl 1 Wojciech Szpankowski 2 Bell Labs Dept. Math.
More informationThe Moments of the Profile in Random Binary Digital Trees
Journal of mathematics and computer science 6(2013)176-190 The Moments of the Profile in Random Binary Digital Trees Ramin Kazemi and Saeid Delavar Department of Statistics, Imam Khomeini International
More informationAnalytic Depoissonization and Its Applications Λ
Analytic Depoissonization and Its Applications Λ W. Szpankowski y Department of Computer Science Purdue University W. Lafayette, IN 47907 May 31, 2003 Λ This research is supported by NSF and NIH. y Joint
More informationAVERAGE-CASE ANALYSIS OF COUSINS IN m-ary TRIES
J. Appl. Prob. 45, 888 9 (28) Printed in England Applied Probability Trust 28 AVERAGE-CASE ANALYSIS OF COUSINS IN m-ary TRIES HOSAM M. MAHMOUD, The George Washington University MARK DANIEL WARD, Purdue
More informationFrom the Discrete to the Continuous, and Back... Philippe Flajolet INRIA, France
From the Discrete to the Continuous, and Back... Philippe Flajolet INRIA, France 1 Discrete structures in... Combinatorial mathematics Computer science: data structures & algorithms Information and communication
More informationThe Subtree Size Profile of Plane-oriented Recursive Trees
The Subtree Size Profile of Plane-oriente Recursive Trees Michael FUCHS Department of Applie Mathematics National Chiao Tung University Hsinchu, 3, Taiwan Email: mfuchs@math.nctu.eu.tw Abstract In this
More informationGrowing and Destroying Catalan Stanley Trees
Discrete Mathematics and Theoretical Computer Science DMTCS vol. 20:1, 2018, #11 Growing and Destroying Catalan Stanley Trees Benjamin Hackl 1 Helmut Prodinger 2 arxiv:1704.03734v3 [math.co] 26 Feb 2018
More informationString Complexity. Dedicated to Svante Janson for his 60 Birthday
String Complexity Wojciech Szpankowski Purdue University W. Lafayette, IN 47907 June 1, 2015 Dedicated to Svante Janson for his 60 Birthday Outline 1. Working with Svante 2. String Complexity 3. Joint
More informationDigital Trees and Memoryless Sources: from Arithmetics to Analysis
Digital Trees and Memoryless Sources: from Arithmetics to Analysis Philippe Flajolet, Mathieu Roux, Brigitte Vallée AofA 2010, Wien 1 What is a digital tree, aka TRIE? = a data structure for dynamic dictionaries
More informationEXPECTED WORST-CASE PARTIAL MATCH IN RANDOM QUADTRIES
EXPECTED WORST-CASE PARTIAL MATCH IN RANDOM QUADTRIES Luc Devroye School of Computer Science McGill University Montreal, Canada H3A 2K6 luc@csmcgillca Carlos Zamora-Cura Instituto de Matemáticas Universidad
More informationPartial Match Queries in Relaxed Multidimensional Search Trees 1
Algorithmica 2001 29: 181 204 DOI: 10.1007/s004530010065 Algorithmica 2001 Springer-Verlag New York Inc. Partial Match Queries in Relaxed Multidimensional Search Trees 1 C. Martínez, 2 A. Panholzer, 3
More informationThe Subtree Size Profile of Bucket Recursive Trees
Iranian Journal of Mathematical Sciences and Informatics Vol., No. (206, pp - DOI: 0.7508/ijmsi.206.0.00 The Subtree Size Profile of Bucket Recursive Trees Ramin Kazemi Department of Statistics, Imam Khomeini
More informationLattice paths with catastrophes
Lattice paths with catastrophes Lattice paths with catastrophes SLC 77, Strobl 12.09.2016 Cyril Banderier and Michael Wallner Laboratoire d Informatique de Paris Nord, Université Paris Nord, France Institute
More informationPROBABILISTIC BEHAVIOR OF ASYMMETRIC LEVEL COMPRESSED TRIES
PROBABILISTIC BEAVIOR OF ASYMMETRIC LEVEL COMPRESSED TRIES Luc Devroye Wojcieh Szpankowski School of Computer Science Department of Computer Sciences McGill University Purdue University 3450 University
More informationApplications of Analytic Combinatorics in Mathematical Biology (joint with H. Chang, M. Drmota, E. Y. Jin, and Y.-W. Lee)
Applications of Analytic Combinatorics in Mathematical Biology (joint with H. Chang, M. Drmota, E. Y. Jin, and Y.-W. Lee) Michael Fuchs Department of Applied Mathematics National Chiao Tung University
More informationMultiple Choice Tries and Distributed Hash Tables
Multiple Choice Tries and Distributed Hash Tables Luc Devroye and Gabor Lugosi and Gahyun Park and W. Szpankowski January 3, 2007 McGill University, Montreal, Canada U. Pompeu Fabra, Barcelona, Spain U.
More information5. Analytic Combinatorics
ANALYTIC COMBINATORICS P A R T O N E 5. Analytic Combinatorics http://aofa.cs.princeton.edu Analytic combinatorics is a calculus for the quantitative study of large combinatorial structures. Features:
More informationAsymptotic results for silent elimination
Discrete Mathematics and Theoretical Computer Science DMTCS vol. subm., by the authors, 1 1 Asymptotic results for silent elimination Guy Louchard 1 and Helmut Prodinger 2 1 Université Libre de Bruxelles,
More informationNotes by Zvi Rosen. Thanks to Alyssa Palfreyman for supplements.
Lecture: Hélène Barcelo Analytic Combinatorics ECCO 202, Bogotá Notes by Zvi Rosen. Thanks to Alyssa Palfreyman for supplements.. Tuesday, June 2, 202 Combinatorics is the study of finite structures that
More informationProbabilistic analysis of the asymmetric digital search trees
Int. J. Nonlinear Anal. Appl. 6 2015 No. 2, 161-173 ISSN: 2008-6822 electronic http://dx.doi.org/10.22075/ijnaa.2015.266 Probabilistic analysis of the asymmetric digital search trees R. Kazemi a,, M. Q.
More information50 years of Linear Probing Hashing
50 years of Linear Probing Hashing Alfredo Viola Universidad de la República, Uruguay AofA 2013 Dedicated to Philippe Flajolet The problem. The table has m places to hash (park) from 0 to m 1, and n elements
More informationAverage Case Analysis of the Boyer-Moore Algorithm
Average Case Analysis of the Boyer-Moore Algorithm TSUNG-HSI TSAI Institute of Statistical Science Academia Sinica Taipei 115 Taiwan e-mail: chonghi@stat.sinica.edu.tw URL: http://www.stat.sinica.edu.tw/chonghi/stat.htm
More informationAnalytic Information Theory: From Shannon to Knuth and Back. Knuth80: Piteaa, Sweden, 2018 Dedicated to Don E. Knuth
Analytic Information Theory: From Shannon to Knuth and Back Wojciech Szpankowski Center for Science of Information Purdue University January 7, 2018 Knuth80: Piteaa, Sweden, 2018 Dedicated to Don E. Knuth
More informationCompositions and samples of geometric random variables with constrained multiplicities
FPSAC 2010, San Francisco, USA DMTCS proc. AN, 2010, 317 328 Compositions and samples of geometric random variables with constrained multiplicities Margaret Archibald 1 and Arnold Knopfmacher 2 and Toufik
More informationCombinatorial/probabilistic analysis of a class of search-tree functionals
Combinatorial/probabilistic analysis of a class of search-tree functionals Jim Fill jimfill@jhu.edu http://www.mts.jhu.edu/ fill/ Mathematical Sciences The Johns Hopkins University Joint Statistical Meetings;
More informationIntroduction Consider a group of n people (users, computers, objects, etc.) sharing a scarce resource (e.g., channel, CPU, etc.). The following elimin
ANALYSIS OF AN ASYMMETRIC LEADER ELECTION ALGORITHM September 3, 996 Svante Janson Wojciech Szpankowski Department of Mathematics Department of Computer Science Uppsala University Purdue University 756
More informationthe electronic journal of combinatorics 4 (997), #R7 2 (unbiased) coin tossing process by Prodinger [9] (cf. also []), who provided the rst non-trivia
ANALYSIS OF AN ASYMMETRIC LEADER ELECTION ALGORITHM Svante Janson Wojciech Szpankowski Department of Mathematics Department of Computer Science Uppsala University Purdue University 756 Uppsala W. Lafayette,
More informationLIMITING DISTRIBUTIONS FOR THE NUMBER OF INVERSIONS IN LABELLED TREE FAMILIES
LIMITING DISTRIBUTIONS FOR THE NUMBER OF INVERSIONS IN LABELLED TREE FAMILIES ALOIS PANHOLZER AND GEORG SEITZ Abstract. We consider so-called simple families of labelled trees, which contain, e.g., ordered,
More informationEverything You Always Wanted to Know about Quicksort, but Were Afraid to Ask. Marianne Durand
Algorithms Seminar 200 2002, F. Chyzak (ed., INRIA, (2003, pp. 57 62. Available online at the URL http://algo.inria.fr/seminars/. Everything You Always Wanted to Know about Quicksort, but Were Afraid to
More informationTHE DEGREE DISTRIBUTION OF RANDOM PLANAR GRAPHS
THE DEGREE DISTRIBUTION OF RANDOM PLANAR GRAPHS Michael Drmota joint work with Omer Giménez and Marc Noy Institut für Diskrete Mathematik und Geometrie TU Wien michael.drmota@tuwien.ac.at http://www.dmg.tuwien.ac.at/drmota/
More informationLIMIT LAWS FOR SUMS OF FUNCTIONS OF SUBTREES OF RANDOM BINARY SEARCH TREES
LIMIT LAWS FOR SUMS OF FUNCTIONS OF SUBTREES OF RANDOM BINARY SEARCH TREES Luc Devroye luc@cs.mcgill.ca Abstract. We consider sums of functions of subtrees of a random binary search tree, and obtain general
More informationMod-φ convergence I: examples and probabilistic estimates
Mod-φ convergence I: examples and probabilistic estimates Valentin Féray (joint work with Pierre-Loïc Méliot and Ashkan Nikeghbali) Institut für Mathematik, Universität Zürich Summer school in Villa Volpi,
More informationTheory and Practice of (some) Probabilistic Counting Algorithms
ICALP-2004, Turku, JUL 2004 Theory and Practice of (some) Probabilistic Counting Algorithms Philippe Flajolet, INRIA, Rocquencourt http://algo.inria.fr/flajolet 1 From Estan-Varghese-Fisk: traces of attacks
More informationConsidering our result for the sum and product of analytic functions, this means that for (a 0, a 1,..., a N ) C N+1, the polynomial.
Lecture 3 Usual complex functions MATH-GA 245.00 Complex Variables Polynomials. Construction f : z z is analytic on all of C since its real and imaginary parts satisfy the Cauchy-Riemann relations and
More informationAnalytic Pattern Matching: From DNA to Twitter. AxA Workshop, Venice, 2016 Dedicated to Alberto Apostolico
Analytic Pattern Matching: From DNA to Twitter Wojciech Szpankowski Purdue University W. Lafayette, IN 47907 June 19, 2016 AxA Workshop, Venice, 2016 Dedicated to Alberto Apostolico Joint work with Philippe
More information4. Complex Analysis, Rational and Meromorphic Asymptotics
ANALYTIC COMBINATORICS P A R T T W O 4. Complex Analysis, Rational and Meromorphic Asymptotics http://ac.cs.princeton.edu ANALYTIC COMBINATORICS P A R T T W O Analytic Combinatorics Philippe Flajolet and
More informationEnumerating connected labelled graphs
Enumerating connected labelled graphs Keith Briggs Keith.Briggs@bt.com more.btexact.com/people/briggsk2/ 2004 August 17 conn-lab-graphs.tex typeset 2004 August 17 10:40 in pdflatex on a linux system Enumerating
More informationNetworks: Lectures 9 & 10 Random graphs
Networks: Lectures 9 & 10 Random graphs Heather A Harrington Mathematical Institute University of Oxford HT 2017 What you re in for Week 1: Introduction and basic concepts Week 2: Small worlds Week 3:
More informationPartial match queries in random quadtrees
Partial match queries in random quadtrees Nicolas Broutin nicolas.broutin@inria.fr Projet Algorithms INRIA Rocquencourt 78153 Le Chesnay France Ralph Neininger and Henning Sulzbach {neiningr, sulzbach}@math.uni-frankfurt.de
More informationNUMBER OF SYMBOL COMPARISONS IN QUICKSORT
NUMBER OF SYMBOL COMPARISONS IN QUICKSORT Brigitte Vallée (CNRS and Université de Caen, France) Joint work with Julien Clément, Jim Fill and Philippe Flajolet Plan of the talk. Presentation of the study
More informationLAWS OF LARGE NUMBERS AND TAIL INEQUALITIES FOR RANDOM TRIES AND PATRICIA TREES
LAWS OF LARGE NUMBERS AND TAIL INEQUALITIES FOR RANDOM TRIES AND PATRICIA TREES Luc Devroye School of Computer Science McGill University Montreal, Canada H3A 2K6 luc@csmcgillca June 25, 2001 Abstract We
More informationHomework 6: Solutions Sid Banerjee Problem 1: (The Flajolet-Martin Counter) ORIE 4520: Stochastics at Scale Fall 2015
Problem 1: (The Flajolet-Martin Counter) In class (and in the prelim!), we looked at an idealized algorithm for finding the number of distinct elements in a stream, where we sampled uniform random variables
More informationOn (Effective) Analytic Combinatorics
On (Effective) Analytic Combinatorics Bruno Salvy Inria & ENS de Lyon Séminaire Algorithmes Dec. 2017 Combinatorics, Randomness and Analysis From simple local rules, a global structure arises. A quest
More informationarxiv: v1 [math.co] 12 Apr 2018
COMBINATORIAL ANALYSIS OF GROWTH MODELS FOR SERIES-PARALLEL NETWORKS MARKUS KUBA AND ALOIS PANHOLZER arxiv:804.0550v [math.co] Apr 08 Abstract. We give combinatorial descriptions of two stochastic growth
More informationCounting Palindromes According to r-runs of Ones Using Generating Functions
Counting Palindromes According to r-runs of Ones Using Generating Functions Helmut Prodinger Department of Mathematics Stellenbosch University 7602 Stellenbosch South Africa hproding@sun.ac.za Abstract
More informationClassification of Join Ordering Problems
Classification of Join Ordering Problems We distinguish four different dimensions: 1. query graph class: chain, cycle, star, and clique 2. join tree structure: left-deep, zig-zag, or bushy trees 3. join
More informationOn the number of heaps and the cost of heap construction
On the number of heaps and the cost of heap construction Hsien-Kuei Hwang Institute of Statistical Science Academia Sinica Taipei 5 Taiwan Jean-Marc Steyaert LIX École Polytechnique 928 Palaiseau France
More informationA New Binomial Recurrence Arising in a Graphical Compression Algorithm
A New Binomial Recurrence Arising in a Graphical Compression Algorithm Yongwoo Choi, Charles Knessl, Wojciech Szpanowsi To cite this version: Yongwoo Choi, Charles Knessl, Wojciech Szpanowsi. A New Binomial
More informationThe space requirement of m-ary search trees: distributional asymptotics for m 27
1 The space requirement of m-ary search trees: distributional asymptotics for m 7 James Allen Fill 1 and Nevin Kapur 1 Applied Mathematics and Statistics, The Johns Hopkins University, 3400 N. Charles
More information6.1 Moment Generating and Characteristic Functions
Chapter 6 Limit Theorems The power statistics can mostly be seen when there is a large collection of data points and we are interested in understanding the macro state of the system, e.g., the average,
More informationI forgot to mention last time: in the Ito formula for two standard processes, putting
I forgot to mention last time: in the Ito formula for two standard processes, putting dx t = a t dt + b t db t dy t = α t dt + β t db t, and taking f(x, y = xy, one has f x = y, f y = x, and f xx = f yy
More informationSingularity Analysis: A Perspective
MSRI, Berkeley, June 2004 Singularity Analysis: A Perspective Philippe Flajolet (Inria, France) Analysis of Algorithms Average-Case, Probabilistic Properties of Random Structures? Counting and asymptotics
More informationAdditive irreducibles in α-expansions
Publ. Math. Debrecen Manuscript Additive irreducibles in α-expansions By Peter J. Grabner and Helmut Prodinger * Abstract. The Bergman number system uses the base α = + 5 2, the digits 0 and, and the condition
More informationn! (k 1)!(n k)! = F (X) U(0, 1). (x, y) = n(n 1) ( F (y) F (x) ) n 2
Order statistics Ex. 4. (*. Let independent variables X,..., X n have U(0, distribution. Show that for every x (0,, we have P ( X ( < x and P ( X (n > x as n. Ex. 4.2 (**. By using induction or otherwise,
More informationDigital search trees JASS
Digital search trees Analysis of different digital trees with Rice s integrals. JASS Nicolai v. Hoyningen-Huene 28.3.2004 28.3.2004 JASS 04 - Digital search trees 1 content Tree Digital search tree: Definition
More informationDISCRETE AND ALGEBRAIC STRUCTURES
DISCRETE AND ALGEBRAIC STRUCTURES Master Study Mathematics MAT40 Winter Semester 015/16 (Due to organisational reasons register for the course at TU and KFU) Edited by Mihyun Kang TU Graz Contents I Lectures
More informationCounting Palindromes According to r-runs of Ones Using Generating Functions
3 47 6 3 Journal of Integer Sequences, Vol. 7 (04), Article 4.6. Counting Palindromes According to r-runs of Ones Using Generating Functions Helmut Prodinger Department of Mathematics Stellenbosch University
More informationThe Expected Profile of Digital Search Trees
The Expected Profile of Digital Search Trees March 24, 2011 Michael Drmota Wojciech Szpankowski Inst. für Diskrete Mathematik und Geometrie Department of Computer Science TU Wien Purdue University A-1040
More informationor E ( U(X) ) e zx = e ux e ivx = e ux( cos(vx) + i sin(vx) ), B X := { u R : M X (u) < } (4)
:23 /4/2000 TOPIC Characteristic functions This lecture begins our study of the characteristic function φ X (t) := Ee itx = E cos(tx)+ie sin(tx) (t R) of a real random variable X Characteristic functions
More informationLECTURE Questions 1
Contents 1 Questions 1 2 Reality 1 2.1 Implementing a distribution generator.............................. 1 2.2 Evaluating the generating function from the series expression................ 2 2.3 Finding
More informationA Master Theorem for Discrete Divide and Conquer Recurrences
A Master Theorem for Discrete Divide and Conquer Recurrences MICHAEL DRMOTA and WOJCIECH SZPANKOWSKI TU Wien and Purdue University Divide-and-conquer recurrences are one of the most studied equations in
More informationLaws of large numbers and tail inequalities for random tries and PATRICIA trees
Journal of Computational and Applied Mathematics 142 2002 27 37 www.elsevier.com/locate/cam Laws of large numbers and tail inequalities for random tries and PATRICIA trees Luc Devroye 1 School of Computer
More informationOn rational approximation of algebraic functions. Julius Borcea. Rikard Bøgvad & Boris Shapiro
On rational approximation of algebraic functions http://arxiv.org/abs/math.ca/0409353 Julius Borcea joint work with Rikard Bøgvad & Boris Shapiro 1. Padé approximation: short overview 2. A scheme of rational
More informationarxiv: v1 [math.co] 1 Aug 2018
REDUCING SIMPLY GENERATED TREES BY ITERATIVE LEAF CUTTING BENJAMIN HACKL, CLEMENS HEUBERGER, AND STEPHAN WAGNER arxiv:1808.00363v1 [math.co] 1 Aug 2018 ABSTRACT. We consider a procedure to reduce simply
More informationMetric Diophantine Approximation for Formal Laurent Series over Finite Fields
Metric Diophantine Approximation for Formal Laurent Series over Finite Fields Michael Fuchs Department of Applied Mathematics National Chiao Tung University Hsinchu, Taiwan Fq9, July 16th, 2009 Michael
More informationAnalytic Algorithmics, Combinatorics, and Information Theory
Analytic Algorithmics, Combinatorics, and Information Theory W. Szpankowski Department of Computer Science Purdue University W. Lafayette, IN 47907 September 11, 2006 AofA and IT logos Research supported
More informationRENEWAL THEORY IN ANALYSIS OF TRIES AND STRINGS: EXTENDED ABSTRACT
RENEWAL THEORY IN ANALYSIS OF TRIES AND STRINGS: EXTENDED ABSTRACT SVANTE JANSON Abstract. We give a survey of a number of simple applications of renewal theory to problems on random strings, in particular
More informationGENERALIZED STIRLING PERMUTATIONS, FAMILIES OF INCREASING TREES AND URN MODELS
GENERALIZED STIRLING PERMUTATIONS, FAMILIES OF INCREASING TREES AND URN MODELS SVANTE JANSON, MARKUS KUBA, AND ALOIS PANHOLZER ABSTRACT. Bona [6] studied the distribution of ascents, plateaux and descents
More informationSORTING ALGORITHMS FOR BROADCAST COMMUNICATIONS: MATHEMATICAL ANALYSIS
SORTING ALGORITHMS FOR BROADCAST COMMUNICATIONS: MATHEMATICAL ANALYSIS PETER J. GRABNER AND HELMUT PRODINGER Abstract. Assume that n persons communicate via a channel; to each person a certain number is
More informationAlmost all trees have an even number of independent sets
Almost all trees have an even number of independent sets Stephan G. Wagner Department of Mathematical Sciences Stellenbosch University Private Bag X1, Matieland 7602, South Africa swagner@sun.ac.za Submitted:
More informationStat 134 Fall 2011: Notes on generating functions
Stat 3 Fall 0: Notes on generating functions Michael Lugo October, 0 Definitions Given a random variable X which always takes on a positive integer value, we define the probability generating function
More informationAsymptotic properties of some minor-closed classes of graphs
FPSAC 2013 Paris, France DMTCS proc AS, 2013, 599 610 Asymptotic properties of some minor-closed classes of graphs Mireille Bousquet-Mélou 1 and Kerstin Weller 2 1 CNRS, LaBRI, UMR 5800, Université de
More informationThe Number of Inversions in Permutations: A Saddle Point Approach
1 2 3 47 6 23 11 Journal of Integer Sequences, Vol. 6 (23), Article 3.2.8 The Number of Inversions in Permutations: A Saddle Point Approach Guy Louchard Département d Informatique CP 212, Boulevard du
More informationOn Buffon Machines & Numbers
On Buffon Machines & Numbers Philippe Flajolet, Maryse Pelletier, Michèle Soria AofA 09, Fréjus --- June 2009 [INRIA-Rocquencourt & LIP6, Paris] 1 1733: Countess Buffon drops her knitting kit on the floor.
More informationScale free random trees
Scale free random trees Tamás F. Móri Department of Probability Theory and Statistics, Eötvös Loránd University, 7 Budapest, Pázmány Péter s. /C moritamas@ludens.elte.hu Research supported by the Hungarian
More informationAscending Runs in Cayley Trees and Mappings
Ascending Runs in Cayley Trees and Mappings Marie-Louise Bruner and Alois Panholzer marie-louise.bruner@tuwien.ac.at alois.panholzer@tuwien.ac.at arxiv:507.05484v [math.co] 20 Jul 205 Vienna University
More informationThe Markov Chain Monte Carlo Method
The Markov Chain Monte Carlo Method Idea: define an ergodic Markov chain whose stationary distribution is the desired probability distribution. Let X 0, X 1, X 2,..., X n be the run of the chain. The Markov
More informationAsymptotics of the Eulerian numbers revisited: a large deviation analysis
Asymptotics of the Eulerian numbers revisited: a large deviation analysis Guy Louchard November 204 Abstract Using the Saddle point method and multiseries expansions we obtain from the generating function
More informationThe Central Limit Theorem: More of the Story
The Central Limit Theorem: More of the Story Steven Janke November 2015 Steven Janke (Seminar) The Central Limit Theorem:More of the Story November 2015 1 / 33 Central Limit Theorem Theorem (Central Limit
More informationPartial Quicksort. Conrado Martínez
Partial Quicksort Conrado Martínez Abstract This short note considers the following common problem: rearrange a given array with n elements, so that the first m places contain the m smallest elements in
More informationof Classical Constants Philippe Flajolet and Ilan Vardi February 24, 1996 Many mathematical constants are expressed as slowly convergent sums
Zeta Function Expansions of Classical Constants Philippe Flajolet and Ilan Vardi February 24, 996 Many mathematical constants are expressed as slowly convergent sums of the form C = f( ) () n n2a for some
More informationarxiv: v2 [math.co] 26 Mar 2016
A new bijection on m-dyck paths with application to random sampling arxiv:603.06290v2 [math.co] 26 Mar 206 Axel Bacher September 5, 208 Abstract We present a new bijection between variants of m-dyck paths
More informationNUMBER OF SYMBOL COMPARISONS IN QUICKSORT AND QUICKSELECT
NUMBER OF SYMBOL COMPARISONS IN QUICKSORT AND QUICKSELECT Brigitte Vallée (CNRS and Université de Caen, France) Joint work with Julien Clément, Jim Fill and Philippe Flajolet Plan of the talk. Presentation
More informationCounting Markov Types
Counting Markov Types Philippe Jacquet, Charles Knessl, Wojciech Szpankowski To cite this version: Philippe Jacquet, Charles Knessl, Wojciech Szpankowski. Counting Markov Types. Drmota, Michael and Gittenberger,
More informationThe number of inversions in permutations: A saddle point approach
The number of inversions in permutations: A saddle point approach Guy Louchard and Helmut Prodinger November 4, 22 Abstract Using the saddle point method, we obtain from the generating function of the
More informationMinimax Redundancy for Large Alphabets by Analytic Methods
Minimax Redundancy for Large Alphabets by Analytic Methods Wojciech Szpankowski Purdue University W. Lafayette, IN 47907 March 15, 2012 NSF STC Center for Science of Information CISS, Princeton 2012 Joint
More informationRapid Design of Subcritical Airfoils. Prabir Daripa Department of Mathematics Texas A&M University
Rapid Design of Subcritical Airfoils Prabir Daripa Department of Mathematics Texas A&M University email: daripa@math.tamu.edu In this paper, we present a fast, efficient and accurate algorithm for rapid
More informationk-protected VERTICES IN BINARY SEARCH TREES
k-protected VERTICES IN BINARY SEARCH TREES MIKLÓS BÓNA Abstract. We show that for every k, the probability that a randomly selected vertex of a random binary search tree on n nodes is at distance k from
More informationASYMPTOTICS OF THE EULERIAN NUMBERS REVISITED: A LARGE DEVIATION ANALYSIS
ASYMPTOTICS OF THE EULERIAN NUMBERS REVISITED: A LARGE DEVIATION ANALYSIS GUY LOUCHARD ABSTRACT Using the Saddle point method and multiseries expansions we obtain from the generating function of the Eulerian
More informationHamming Weight of the Non-Adjacent-Form under Various Input Statistics. Clemens Heuberger and Helmut Prodinger
FoSP Algorithmen & mathematische Modellierung FoSP Forschungsschwerpunkt Algorithmen und mathematische Modellierung Hamming Weight of the Non-Adjacent-Form under Various Input Statistics Clemens Heuberger
More informationDerangements with an additional restriction (and zigzags)
Derangements with an additional restriction (and zigzags) István Mező Nanjing University of Information Science and Technology 2017. 05. 27. This talk is about a class of generalized derangement numbers,
More information