50 years of Linear Probing Hashing

Size: px
Start display at page:

Download "50 years of Linear Probing Hashing"

Transcription

1 50 years of Linear Probing Hashing Alfredo Viola Universidad de la República, Uruguay AofA 2013 Dedicated to Philippe Flajolet

2 The problem. The table has m places to hash (park) from 0 to m 1, and n elements (cars). Each element is given a hash value (preferred parking lot). If place is empty, then the element is stored there. Otherwise, looks sequentially for an empty place. If no empty place up to the end of the table, the search follows at location 0. Several R.V. to study, mainly related with cost of individual searches and total construction cost. Very important special case: Parking Problem. In parking the car is lost if no empty place up to the end. Main R.V. is the number of lost cars. Important variant: each location can hold up to b (or k) cars.

3 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

4 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

5 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

6 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

7 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

8 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

9 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

10 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

11 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

12 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

13 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

14 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

15 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

16 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

17 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

18 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, / 37

19 A discrete parking problem Limiting distribution results Analysis Further research A discrete parking problem: Example Example: 8 parking lots, 8 cars Parking sequence: 3, 6, 3, 8, 6, 7, 4, cars are unsuccessful 5 / 37

20 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

21 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

22 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

23 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

24 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

25 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

26 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

27 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

28 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

29 A discrete parking problem Limiting distribution results Analysis Further research Further research: Bucket parking scheme Blake and Konheim [1976]: Bucket parking scheme Each parking lots can hold up to r cars Related to analysis of bucket hashing algorithms r 32 / 37

30 Linear Probing Hashing.

31 The mathematical beauty of Linear Probing!

32 Some personal motivations... Simplest collision resolution strategy for open addressing [Peterson 1957]. Works well for tables that are not too full. Because of primary clustering, its performance deteriorates when the load factor is high. Its analysis leads to nontrivial and interesting mathematical problems. There are connections with tree inversions, tree path lenghts, graph connectivity, area under excursion, etc. Equivalent formulation in terms of the parking problem. First problem that D. Knuth analyzed [Knuth 1962] with bucket size 1, and motivated the collection "The Art of Computer Programming". The analysis for general bucket size b presents very interesting challenges. For example, can symbolic methods be used?

33 ... and a box full of surprises... The study of parking sequences and their deep relations with other problems in both discrete and continuous mathematics has been carried out by dierent research communities in parallel and with little communication among them. Several related problems have been studied by experts in probability, combinatorics and computer sciences. The methodological techniques to study these problems are very diverse, and cover a wide range of research areas. As it is said in [Chassaing and Flajolet 2003], a systematic historical approach to the problem is very dicult. I will concentrate on the contributions made by people from our community as well as some related work that have inspired their research.

34 1962: Summer work by Don Knuth...

35 First published paper.

36 Historical note on rst published paper.

37 Original results. Let a hash table with m positions and n inserted elements. Let P m;n the probability of the last position being empty. P m;n = 1 n m. Let C m;n the R.V. for the number of successful searches of a random element. E [C m;n ] = 1 2 (1 + Q 0((m; n 1)). E [C m;m ] = with 0 < 1. p E [C n n;n ] = + O(1) (proved on may 20, 1965). 8 Let U m;n the R.V. for the number of unsuccessful searches of a random element. E [U m;n ] = 1 2 (1 + Q 1((m; n 1)). E [U m;m ] = (1 ) 2 with 0 < 1. The Ramanujan Q function is the special case Q 0 (n; n) of! nx k + r n k Q r (m; n) = k m n : k=0

38 ... question by Ramanujan to Hardy in 1913.

39 The problem...

40 ... and the Ramanujan's Q function into play!

41 Methodology...

42 ... and more methodology!

43 Collision Resolution Strategies. In open addressing, when two keys collide, either one of them may stay in that location, while the other one keeps probing. First-Come-First-Served (standard). Each collision is resolved in favor of the rst record that probed the location. Last-Come-First-Served [Poblete and Munro 1989]. Each collision is resolved in favor of the incoming record. Robin Hood [Celis, Larson and Munro 1985]. Each collision is resolved in favor of the record that is further away from its home location.

44 Original proposal of Robin Hood. "Analysis of a le addressing method" [Schay and Spruth 1962]. Modication of the algorithm proposed by Peterson in This is the Robin Hood strategy! Expected value of search cost does not change, as already noted by D. Knuth in his original note. They do an analysis based on a Poisson approximation, and nd E [C m;m ] = with 0 < 1.

45 Two models to analyze the problem. Exact lling model. A xed number of keys n, are distributed among m locations, and all m n possible arrangements are equally likely to occur. Poisson model. Each location receives a number of keys that is Poisson distributed with parameter b, and is independent of the number of keys going elsewhere. This implies that the total number of keys, N, is itself a Poisson distributed random variable with parameter bm: P r[n e bm (bm) n = n] = : n!

46 Poisson Transform. Results in one model can be transfered into the other model by the Poisson Transform: X X P m [f m;n ; b] = P r[n = n]f m;n = e bm (bm) n f m;n : n! n0 n0 Inversion Theorem: [Gonnet and Munro 1984] If P m [f m;n ; b] = X k0 a m;k (bm) k then f m;n = X k0 a m;k n k (bm) k : The Poisson model is an approximation of the exact lling model when n; m! 1 with n=m = b with 0 < 1.

47 , Diagonal Poisson Transform. [Munro, Poblete, Viola 1997] Let a hash table of size m, with n + 1 keys, and let P be a property for (chosen uniformily at random). Let f m;n be the result of applying a linear operator (e.g. an expected value) to the probability generating function of P. X f m;n = Pr[ 2 cluster of size i + 1] f i+2;i i0 X n (m i 2) n i 1 (m n 2)(i + 2) i = i m n f i+2;i : i0 Then P m [f m;n ; ] = D 2 [f n+2;n ; ] with D c [f n ; ] = (1 ) X n0 e (n+c) ((n + c))n f n : n!

48 Combinatorial interpretation. Any Linear Probing Hash table can be seen as a sequence of almost full tables (a subtable with all but the last bucket full). Example: [3-3],[4-4],[5-5],[6-2]. This interpretation can be nicely handled by Analytic Combinatorics, since for example, it implies that it is enough to study almost full tables, and then use the sequence construction.

49 Distribution of individual displacements. Complementary probabilistic and combinatorial approaches. [Janson 2005, Viola 2005]. Let P (z) be the probability generating function for the displacement of a random element wth 0 < 1 and 2 ff CF S; RHg. Then, with T (x) = ze T (x) the tree function, P F CF S (z) = (1 T (ze )) 2 (1 ) 2 2(1 z) P RH (z) = 1 e z e ze e z : LCFS is much more challenging and [Janson 2005] presents an exact expression. Exact results follow by depoissonization. :

50 Parking sequences, acyclic maps, priority queues. [Seitz 2009] surveys some of these relations. [Gilbey and Kalikow 1999]: priority queues. Presents generalizations of the parking problem. Distributional analysis of the overow. Analysis of the overow in parking with buckets. g(m; n; k): the number of defective parking functions of defect k. Let G(z; u; v) = Then X X X m0 n0 d0 G(z; u; v) = 1 g(m; n; k) zn n! um v d. T (zu) 1 zv T (zu) z 1 u ezv : v [Cameron, Johannsen, Prellberg and Schweitzer 2008]. In [Viola 2005] parking is used as a subproblem for RH.

51 Parking problem: limit distributions [Panholzer 2008] Studies P rfx m;n = kg = g(m;n;k) m n. Limit distributions for X m;n. Nine regions depending on growth of m; n. n m: X m;n L! X with P rfx = 0g = 1 (degenerated law). n m; 0 < < 1: X L m;n! X, discrete limit law. p m := m n m: m X m;n L! X(d) = EXP(2). := m n p m; > 0: 0 := m n p m: 1 p m X m;n L! X (d) = LINEXP(2,). 1 p m X m;n L! X (d) = RAYLEIGH(2). 0 := n m p n: Xm;n+m p n n X L m;n! X (d) = RAY(2). := n m p n; > 0: Xm;n+m n p n X m;n L! X (d) = LE(2,). p n := n m n: m (X m;n + m n) L! X(d) = EXP(2). n m; > 1: (X m;n + m n) L! X, discrete limit law. m n: X m;n L! X with P rfx = 0g = 0 (degenerated law).

52 Total displacement (box full of surprises!).

53

54 4 Analytic Combinatorics Two basic principles 7! \dictionaries" SYMBOLIC METHODS Generating functions 7! z 11 z + z 2 + z z z z z z 8 + Analytic functions and singularities 10

55 CONSTRUCTIONS Dictionary (I) F 7! ff n g 7! f(z) = X n 1 1 f = 1 + f + f 2 + f 3 + f n z n n! : exp(f) = 1 + f + 1 2! f ! f 3 + A[B 7! A(z)+B(z) AB 7! A(z)B(z) Seq A 7! 1 1 A(z) Set A 7! exp(a(z)) Cycle A 7! 1 log 1 A(z) 11

56 COMPLEX ASYMPTOTICS Dictionary (II) Point of regularity. f(z) f(z 0 ) + f 0 (z 0 )(z z 0 ) exp(z) Point of singularity. :9 d dz f(z) z0 (1 z) 7! n 1 () p 1 z 7! n 3=2 (1 z) 3=2 7! n 5=2 12

57 Permutations: (1 z) x y Trees: 1 p 1 z y x

58 Distributional analysis (almost full tables) Record construction cost = total displacement F n (q) = n 1 X k=0 n 1 k F k (q)(1 + q + + q k )F n 1 k (q): q{calculus X n q n2 z 2 The construction cost is worst-case quadratic 17

59 Lemma. The generating functions associated with moments f r (z) = q r F (z; q) satisfy a linear ODE that involves T (z). zf 1 = 1 2 T 3 (1 T ) 2 ; zf 2 = 1 T 4 (24 11T + 2T 2 ) 12 (1 T ) 5 Theorem. [Full tables, exact form of moments] E[d n;n ] = n (Q(n) 1) 2 E[d 2 n;n] = n 12 (5n2 + 4n 1 8n Q(n)): Q(n) := 1 + n 1 (n 1)(n 2) (n 1)(n 2)(n 3) + n n n 3 Theorem. [Full tables, asymptotic form of moments] E[d n;n ] = Var[d n;n ] = p 2 4 n3=2 2 p 2 3 n + 48 n1= O(n 1 ); p n 3 + n n3=2 Knuth (1962-3); Flajolet-Poblete-Viola (1997); Knuth (1997) 20

60 6 Limit distribution A method of \pumping" moments Start from nonlinear combinatorial decomposition (BGF) [F (z; q)] = 0 Apply derivatives U@ r q to get rth moment. Expect linear operator L with Lf r = r [f 0 ; f 1 ; : : : ; f r 1 ] Solve exactly and/or or asymptotically (singularities) Method used on Quicksort, Hennequin 1989: 100 moments; nongaussian law Path length in trees, Takacs Area below walks, Louchard 1984 In situ permutation, Knuth 1972, Prodinger et alii Path length in Cayley trees: F (z; q) ze F (qz;q) = 0 21

61 Moment problem A classical theorem: if the \Moment generating function" M(z) = X r r z r r! has nonzero radius of convergence, then the law is uniquely determined by its moments. A corollary: Convergence of moments implies convergence of distributions Example. w(x) = e x, r r! =) M(z) = 1 1 z Theorem. For almost full tables, convergence to the Airy distribution, Prf d n;n 1 xg! PrfX xg (n! 1); (n=2) 3=2 where X is Airy distributed. E[X r ] = ( 1 2 ) ( 3r 1 2 ) r. X w r r = 2=3(w) r! 1=3(w) r0 (w) = 1 (4 2 1) w + (42 1)(4 2 9) w ! 24 (42 1)(4 2 9)(4 2 25) w 3 + : 3! 24 24

62 The Airy density (area under excursions). [Louchard 1984],[Takács 1991] The constants k satisfy The density!(x) has been calculated in [Takács 1991]. The constants k are the zeros of the Airy function Ai(z).

63 7 Sparse tables A table with m cells and n elements has g = m n \gaps". SparseTable := <Full> *... * <Full> Bivariate GF is: (F (z; q)) g The analysis can be \recycled" (---- g times ----) Theorem. For -sparse tables, = n, mean and variance: m E[dm;n] = n 2 (Q 0 E[d 2 m;n ] = 12 n (m; n 1) 1); (m n) 3 + (n + 3)(m n) 2 + (8n + 1)(m n) + 5n 2 + 4n 1 ((m n) 3 + 4(m n) 2 + (6n + 3)(m n) + 8n)Q 0 (m; n 1) : Q0(m; n) := 1 + n 1 (n 1)(n 2) (n 1)(n 2)(n 3) + m m 2 + m 3 + E[d m;n ] = 2(1 ) n 2(1 ) + 3 O(n 1 ); Var[d m;n ] = (1 ) 4 n Flajolet-Poblete-Viola (1997); Knuth (1997) 26

64 The limit distribution Theorem. A Gaussian law. Proof. Integral of large powers by saddle point [z n ](F (z; q)) m n = 1 2i I m n dz (F (z; q)) z n+1 1e-05 8e-06 6e-06 4e-06 2e y x Plus continuity theorem for characteristic functions q = e i Corollary. Works for large assemblies! E.g. [Mahmoud] Distribution sorts with O(n) buckets. 27

65 Linear Probing and random graphs (I).

66 Mails exchanged with Don Knuth. Date: Mon, 29 Sep :15: (PDT)... To: Subject: note from Don Knuth Dear Ph, Ordinarily I am not happy to receive , but in this case it was very touching to learn that you had decided to dedicate such a nice paper to me, just after I had (secretly) decided to dedicate reference [22] to you! But I haven't time to study it in detail now, as I'm working 150% time on the new edition of Volume Best regards, Don

67 Combinatorial approach to Linear Probing.

68 Combinatorial Analysis (FPV).

69 Solution to the fundamental recurrence (Knuth).

70 Conclusions (Knuth).

71 Properties of the parking function F n (q). Several combinatorial relations between very important combinatorial problems. [Kreweras 1980]

72 Parking, random graphs and random trees. [Spencer 1997]. BFS traversal of a random graph with vertices {0, 1,..., n}. Induces a queue (Q k ( )) 1kn and a spanning tree. BFS induces a parking sequence (A k ( )) 1kn. Ex: (A k ( )) = ff6; 8g; f2; 3g; ; f7g; f1; 4g; f5g; f9g; ; g. x k ( ) = ja k ( )j. Ex: (x k ( )) = f2; 2; 0; 1; 2; 1; 1; 0; 0g. y k ( ) = x 1 ( ) + x 2 ( ) + : : : + x k ( ) k + 1, size of queue (Q( )) before step k. Ex: (y k ( )) = f2; 3; 2; 2; 3; 3; 3; 2; 1g. y 1 ( ) + : : : y n ( ) n is the total displacement. Ex: 12. BFS induces a random walk excursion. Ex: b.

73 Parking sequences and random graphs. C n;k : # connected graphs with n vertices and n + k 1 edges. D n+1;n : RV for total displacement in parking with n cars. Theorem: C n;k C n;0 = E h Dn;n 1 i. k Sketch: How may graphs give the same with this BFS? Graph has the edges of plus some of the (y 1 ( ) 1)+ : : : + (y 1 ( ) 1) edges joining parked car with collisions. C n+1;k = X y 1( )+:::+y 1( ) n k h Dn+1;n i = E k (n + 1) n 1. (n + 1) n 1 is both the number of parking functions with n cars and labelled trees with n + 1 nodes.

74 Back to Knuth (1997). P.g.f. for the total displacement in parking with n cars: F n (q) F n (1) : Moreover and "!# Dn+1;n E = 1 F n (k) (1) k k! F n (1) ; F (1 + q) = X k0 1 k! F n (k) (1)q k : So, if F n (q) enumerates total displacements in parking with n cars, then F n (1 + q) enumerates connected graphs with n + 1 nodes discriminated by their excess!.

75 Inversions in Cayley trees. [Gessel and Wang 1979] The inversions are f4; 3g; f4; 2g; f6; 2g, and f6; 5g. Bijection with parking problem. Same functional equation. BFS starting at 1, visiting the greatest unvisited node rst. How many graphs share the same spanning tree?: Inversions! I n (t): # of inversions of a tree rooted at 1. C n (t) = t n 1 I n (1 + t).

76 Parking sequences, random excursions,... [Spencer 1997]. (A k ( )) = ff6; 8g; f2; 3g; ; f7g; f1; 4g; f5g; f9g; ; g. x k ( ) = ja k ( )j. Ex: (x k ( )) = f2; 2; 0; 1; 2; 1; 1; 0; 0g. y k ( ) = x 1 ( ) + x 2 ( ) + : : : + x k ( ) k + 1, size of queue (A( )) before step k. Ex: (y k ( )) = f2; 3; 2; 2; 3; 3; 3; 2; 1g. n! There are x 1!:::x n! parkings seqs. Q associated to (y 1 ; : : : ; y k ), n e with probability proportional to 1 i=1 x i, (n Po(1) i.i.d.).! Poisson R.W. with y 0 = 0, steps x i 1, conditioned to be positive on time 1; 2; : : : ; n and zero on time n + 1.

77 ... and the ineable Brownian excursion! (x k ( )) 0kn are Po(1) i.i.d. conditioned as above. Corresponding unlabelled tree is a Galton-Watson tree with Po(1) progeny, constrained to have n + 1 nodes, and x k is the progeny of k th node visited by the BFS. ybntc It is known that pn L! (e(t)) 0t1, ( L! denotes 0t1 convergence in law) with (e(t)) the Brownian excursion. R As a consequence Dn+1;n n p L 1 n! 0 e(t)dt: max k y k pn L! m = max e(t): 0t1 [Chassaing and Marckert 2001]. Convergence of moments. Coupling labeled trees-empirical processes using parking functions. Alternative probabilistic proof to [Flajolet, Poblete, Viola 1998] for the convergence of moments towards moments of the Airy law.

78 More asymptotic distributions: new cases. [Janson 2001] Four dierent methods used in the proof! There is a phase transition around m n p m. Let U m;n the average cost of an unsuccessful search beginning at a random cell h (averaged over h). Then

79 "Monkey Saddle" as the ALGO project's logo.

80 Distribution of lengths of parking blocks. [Pittel 1985]. Table with ` = m n empty places. B m;` k length of k th largest cluster. When `=m! with > 0 then B m;` 2 log m 3 log log m+ m 1 = ; where 2( 1 log ) m converges weakly to an extreme-value distribution. Moreover, from [Chassaing and Louchard 2002] we have Theorem: For n; m jointly to +1, 1 if p m = o(`); B m;` 1 =m L! 0; 2 if ` = o( p m); B m;` 1 =m L! 1: Phase transition occurs when ` = ( p m). Largest block reaches O(m) after p m cars arrive at the critical region. So, we have a coalescence phenomena. Studied with the help of Brownian motion theory.

81 Size of the rst created cluster. [Chassaing and Louchard 2002]. R m;n, size of the cluster of the rst arrived car. P r[r m;n = k] = n 1 (k+1) k 1 (m k 1) n k 1 (m n 1) k 1 m n. 1 P r[r m;n = k] (k + 1) k 1 e (k+1) k 1 (1 ), (0 < < 1). P r[r m;n = k] = 1 m f ; k m where f is the density of N 2 2 +N 2 (k 1)! f (; x) = 2 x 1=2 (1 x) 3=2 exp + O(m 3=2 ); m n = p m, (N standard Gaussian), with 1 ]0;1[ (x). 2 x 2(1 x) In other words, when m n = p m then Rm;n m L! : N 2 Limit cases: Rm;n m 2 +N 2. P Rm;n! 0 (! 1) and m P! 1 (! 0).

82 Distribution of the length of the clusters. B m;` k length of the k th largest cluster and B m;` = B m;` k Let e(t) the normalized Brownian excursion. sup Let e(x) = e(x) x 0yx (e(y) y). B() = (B k ()) k1 decreasing widths of e(x) excusrions. Theorem: If lim ` p m = > 0, then Bm;l m L! B(). k1. As a consequence, from [Pavlov 1977] in random forest (and its relation with the parking problem), Joint distribution of B() out of reach.

83 Asymptotic distributions of block lenghts.

84 Size-biased permutations R(). R() constructed from B(). 1 Choose R 1 () with P r(r 1 () = B k ()jb()) = B k (). 2 Same for R k () but with the terms that did not appear before. Let R m;` = R() appears in the -valued fragmentation process derived from the continuum random tree (CRT) in the standard additive coalescence. [Aldous and Pitman 1998]. R m;` k k1, the sequence of block sizes ordered by Rm;` k date of birth, and R k () = m with ` = n m = dp me. p ` m = > 0, then Rm;l L m! R(). Theorem: If lim

85 Coalescence: emergence of a giant block. Up to now parking frozen at xed : n() = m b p mc. To understand coalescence: understand the dependence between parking schemes at times n( 1 ) < : : : < n( k ). Joint distribution: (size(), initial position()) 0 of each block. Coalescence in the discrete model, translates into coalescence in the continuous model at the limit. Two constructions of the additive coalescent. [Aldous and Pitman 1998] CRT as a limit of a discrete model of coalescence-fragmentation process that starts with a random unrooted labelled tree (reverse of process of deleting edges at random). [ Bertoin 2000] based on xcursions of the family of stochastic processes ( e) 0. [Chassaing and Louchard 2002] prove that asymptotically parking schemes lead to this construction of the additive coalescent.

86 Parking schemes and Pavlov's forests.

87 Parking with buckets: distribution of overow. Let g b (m; n; d) the number of parking functions of defect d, with buckets of size b. Let G b (z; u; v) = Then [Seitz 2009] G b (z; u; v) = X X X m0 n0 d0 1 1 u v k e zv g b (m; n; d) zn n! um v d.! b Y1 j=1 b Y1 j=1 1 b 1 b zv T (!j u 1 b z b z T (!j u 1 b z b ) : ) Let w m;b;k be the probability of having k cars going to overow in a b-full table with m buckets of size b and < 1, and m (b; z) = P k0 w m;b;kz k. Then [Viola 2010] Q b 1 Q j=1 z b e b(z 1) b 1 b(1 )(z 1) m (b; z) = j=1 z 1 T (! j e ) T (! j e ) + O bm :

88 Linear Probing with Buckets: bucket occupancy. Let T d (b) be the probability that a given bucket has more than d empty places, when n; m! 1, 0 n=bm = < 1. From [Viola 2010], inspired in [Blake and Konheim 1977]: Theorem Q [u d b 1 ] i=1 T d (b) = b(1 ) Q b 1 j=1 1 1 u T (!i e ) ; 0 d b 1; T (! j e ) where T is the Tree function and! is a b-th root of unity. The sequence T k;d;b = k![ k ]T d (); 0 d < b is sequence EIS A Neil Sloane's Encyclopedia of Integer Sequences.

89 Linear Probing and Paking problem with Buckets. From [Viola 2010], following [Viola and Poblete 1998] and the pioneering work by [Blake and Konheim 1977]: Theorem Let m;b be the random variable for the cost of searching a random element in a b-full table with m buckets of size b and < 1, using the Robin Hood linear probing hashing algorithm, and let m (b; z) be its probability generating function. Then m (b; z) = z b C m (b; z) = b X1 d=0 C m b; e 2id X b z 1=b b 1 p=0 Q 1)) b 1 b(1 )(1 eb(z b z b e b(z 1) Q j=1 b 1 j=1 e 2id b z 1=b p ; with z 1 T (r j e ) T (r j e ) :

90 Some nal considerations. Problem with a very rich history. Paradigm of a problem that nicely integrates analytical, combinatorial and probabilistic approaches. This integration (together with the use of symbolic methods!) has allowed the understanding of deep relations with other important problem. Problem that it is continuously evolving, and new important results are still to come. Some ongoing work: Total displacement with buckets. Relation with other problems as for b=1? FCFS with buckets. Study more properties of sequence T d (b) and nd other applications. Number of movements in deletion algorithm. Worst case Robin Hood.

A unied approach to linear probing hashing

A unied approach to linear probing hashing A unied approach to linear probing hashing Svante Janson Uppsala University Alfredo Viola Universidad de la República AofA 2014 Dedicated to Philippe Flajolet A discrete parking problem Limiting distribution

More information

INDIVIDUAL DISPLACEMENTS FOR LINEAR PROBING HASHING WITH DIFFERENT INSERTION POLICIES

INDIVIDUAL DISPLACEMENTS FOR LINEAR PROBING HASHING WITH DIFFERENT INSERTION POLICIES INDIVIDUAL DISPLACEMENTS FOR LINEAR PROBING HASHING WITH DIFFERENT INSERTION POLICIES SVANTE JANSON Abstract. We study the distribution of the individual displacements in hashing with linear probing for

More information

On (Effective) Analytic Combinatorics

On (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 information

BT Research at Martlesham, Suffolk

BT Research at Martlesham, Suffolk Connectivity of random graphs Keith Briggs Keith.Briggs@bt.com research.btexact.com/teralab/keithbriggs.html Alan Turing Institute 2004 November 19 1330 typeset 2004 November 17 10:55 in pdflatex on a

More information

From the Discrete to the Continuous, and Back... Philippe Flajolet INRIA, France

From 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 information

Approximate Counting via the Poisson-Laplace-Mellin Method (joint with Chung-Kuei Lee and Helmut Prodinger)

Approximate Counting via the Poisson-Laplace-Mellin Method (joint with Chung-Kuei Lee and Helmut Prodinger) 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

More information

Enumerating connected labelled graphs

Enumerating 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 information

5. Analytic Combinatorics

5. 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 information

Notes by Zvi Rosen. Thanks to Alyssa Palfreyman for supplements.

Notes 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 information

Erdős-Renyi random graphs basics

Erdős-Renyi random graphs basics Erdős-Renyi random graphs basics Nathanaël Berestycki U.B.C. - class on percolation We take n vertices and a number p = p(n) with < p < 1. Let G(n, p(n)) be the graph such that there is an edge between

More information

The Subtree Size Profile of Plane-oriented Recursive Trees

The 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 information

Algorithmic Tools for the Asymptotics of Linear Recurrences

Algorithmic Tools for the Asymptotics of Linear Recurrences Algorithmic Tools for the Asymptotics of Linear Recurrences Bruno Salvy Inria & ENS de Lyon Computer Algebra in Combinatorics, Schrödinger Institute, Vienna, Nov. 2017 Motivation p 0 (n)a n+k + + p k (n)a

More information

Forty years of tree enumeration

Forty 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 information

Algorithmica 2001 Springer-Verlag New York Inc.

Algorithmica 2001 Springer-Verlag New York Inc. Algorithmica (001 1: 61 77 DOI: 10.1007/s0045-001-0056-0 Algorithmica 001 Springer-Verlag New York Inc. Analytic Variations on the Airy Distribution 1 P. Flajolet and G. Louchard Abstract. The Airy distribution

More information

Notes on combinatorial graph theory

Notes on combinatorial graph theory Notes on combinatorial graph theory Keith Briggs Keith.Briggs@bt.com more.btexact.com/people/briggsk2/cgt.html CRG meeting 2004 Jan 19 Monday 15:00 comb-graph-th.tex typeset 2004 January 27 11:16 in pdflatex

More information

LIMITING DISTRIBUTIONS FOR THE NUMBER OF INVERSIONS IN LABELLED TREE FAMILIES

LIMITING 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 information

Lattice paths with catastrophes

Lattice 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 information

Combinatorial/probabilistic analysis of a class of search-tree functionals

Combinatorial/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 information

The range of tree-indexed random walk

The range of tree-indexed random walk The range of tree-indexed random walk Jean-François Le Gall, Shen Lin Institut universitaire de France et Université Paris-Sud Orsay Erdös Centennial Conference July 2013 Jean-François Le Gall (Université

More information

Polynomials over finite fields. Algorithms and Randomness

Polynomials over finite fields. Algorithms and Randomness Polynomials over Finite Fields: Algorithms and Randomness School of Mathematics and Statistics Carleton University daniel@math.carleton.ca AofA 10, July 2010 Introduction Let q be a prime power. In this

More information

ADDITIONAL NOTES. 3 Surjections 6

ADDITIONAL NOTES. 3 Surjections 6 Contents Labelling Atoms. Well-labelling............................................. The Exponential Generating Function egf............................3 Examples................................................

More information

The Subtree Size Profile of Bucket Recursive Trees

The 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 information

Random trees and branching processes

Random trees and branching processes Random trees and branching processes Svante Janson IMS Medallion Lecture 12 th Vilnius Conference and 2018 IMS Annual Meeting Vilnius, 5 July, 2018 Part I. Galton Watson trees Let ξ be a random variable

More information

Singularity Analysis: A Perspective

Singularity 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 information

Networks: Lectures 9 & 10 Random graphs

Networks: 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 information

Dependence 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) 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 information

PARKING FUNCTIONS. Richard P. Stanley Department of Mathematics M.I.T Cambridge, MA

PARKING FUNCTIONS. Richard P. Stanley Department of Mathematics M.I.T Cambridge, MA PARKING FUNCTIONS Richard P. Stanley Department of Mathematics M.I.T. -75 Cambridge, MA 09 rstan@math.mit.edu http://www-math.mit.edu/~rstan Transparencies available at: http://www-math.mit.edu/~rstan/trans.html

More information

Counting terms in the binary lambda calculus

Counting terms in the binary lambda calculus Counting terms in the binary lambda calculus Katarzyna Grygiel Theoretical Computer Science Department Jagiellonian University in Kraków Pierre Lescanne Laboratoire de l'informatique du Parallélisme École

More information

THE CENTER OF MASS OF THE ISE AND THE WIENER INDEX OF TREES

THE CENTER OF MASS OF THE ISE AND THE WIENER INDEX OF TREES Elect. Comm. in Probab. 9 (24), 78 87 ELECTRONIC COMMUNICATIONS in PROBABILITY THE CENTER OF MASS OF THE ISE AND THE WIENER INDEX OF TREES SVANTE JANSON Departement of Mathematics, Uppsala University,

More information

Algorithmic Tools for the Asymptotics of Linear Recurrences

Algorithmic Tools for the Asymptotics of Linear Recurrences Algorithmic Tools for the Asymptotics of Linear Recurrences Bruno Salvy Inria & ENS de Lyon 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

More information

Analytic Combinatorics: A Primer

Analytic Combinatorics: A Primer Analytic Combinatorics: A Primer Conrado Martínez Univ. Politècnica de Catalunya, Spain April 2011 What is this course about? Analytic Combinatorics seeks to develop mathematical techniques that help us

More information

ANALYSIS OF SOME TREE STATISTICS

ANALYSIS OF SOME TREE STATISTICS ANALYSIS OF SOME TREE STATISTICS ALOIS PANHOLZER Abstract. In this survey we present results and the used proof techniques concerning the analysis of several parameters in frequently used tree families.

More information

An Analysis of Top Trading Cycles in Two-Sided Matching Markets

An Analysis of Top Trading Cycles in Two-Sided Matching Markets An Analysis of Top Trading Cycles in Two-Sided Matching Markets Yeon-Koo Che Olivier Tercieux July 30, 2015 Preliminary and incomplete. Abstract We study top trading cycles in a two-sided matching environment

More information

Growing and Destroying Catalan Stanley Trees

Growing 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 information

Applications 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) 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 information

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) 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 information

Brownian Motion Area with Generatingfunctionology

Brownian Motion Area with Generatingfunctionology Brownian Motion Area with Generatingfunctionology Uwe Schwerdtfeger RMIT University/University of Melbourne Alexander von Humboldt Foundation Monash University 3 August, 2011 Some continuous time processes...

More information

Descents in Parking Functions

Descents in Parking Functions 1 2 3 47 6 23 11 Journal of Integer Sequences, Vol. 21 (2018), Article 18.2.3 Descents in Parking Functions Paul R. F. Schumacher 1512 Oakview Street Bryan, TX 77802 USA Paul.R.F.Schumacher@gmail.com Abstract

More information

Gibbs distributions for random partitions generated by a fragmentation process

Gibbs distributions for random partitions generated by a fragmentation process Gibbs distributions for random partitions generated by a fragmentation process Nathanaël Berestycki and Jim Pitman University of British Columbia, and Department of Statistics, U.C. Berkeley July 24, 2006

More information

Multiple Choice Tries and Distributed Hash Tables

Multiple 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 information

Deterministic edge-weights in increasing tree families.

Deterministic edge-weights in increasing tree families. Deterministic edge-weights in increasing tree families. Markus Kuba and Stephan Wagner Institut für Diskrete Mathematik und Geometrie Technische Universität Wien Wiedner Hauptstr. 8-0/04, 040 Wien, Austria

More information

Some Node Degree Properties of Series-Parallel Graphs Evolving Under a Stochastic Growth Model

Some Node Degree Properties of Series-Parallel Graphs Evolving Under a Stochastic Growth Model Some Node Degree Properties of Series-Parallel Graphs Evolving Under a Stochastic Growth Model Hosam M. Mahmoud The George Washington University Washington, DC 20052, USA 24th International Meeting on

More information

Some problems in the statistical mechanics of polymers

Some problems in the statistical mechanics of polymers Some problems in the statistical mechanics of polymers Universität Bielefeld Bielefeld, Germany Eindhoven, June 21, 2007 Motivation viewpoint: combinatorics and asymptotic enumeration compute limit laws

More information

Searching with Dice A survey on randomized data structures

Searching with Dice A survey on randomized data structures Searching with Dice A survey on randomized data structures Conrado Martínez Univ. Politècnica de Catalunya, Spain May 11th, 2010 Journée inaugurale de l'équipe Combinatoire, algorithmique et interactions

More information

Unlabelled Structures: Restricted Constructions

Unlabelled Structures: Restricted Constructions Gao & Šana, Combinatorial Enumeration Notes 5 Unlabelled Structures: Restricted Constructions Let SEQ k (B denote the set of sequences of exactly k elements from B, and SEQ k (B denote the set of sequences

More information

COUNTING WITH AB-BA=1

COUNTING WITH AB-BA=1 COUNTING WITH AB-BA=1 PHILIPPE FLAJOLET, ROCQUENCOURT & PAWEL BLASIAK, KRAKÓW COMBINATORIAL MODELS OF ANNIHILATION-CREATION ; IN PREP. (2010) 1 A. = + + B. = 2 Annihilation and Creation A = Annihilate

More information

arxiv: v1 [math.co] 30 Aug 2017

arxiv: v1 [math.co] 30 Aug 2017 Parking Cars of Different Sizes arxiv:1708.09077v1 [math.co] 30 Aug 2017 Richard Ehrenborg and Alex Happ Abstract We extend the notion of parking functions to parking sequences, which include cars of different

More information

Asymptotic Enumeration of Compacted Binary Trees

Asymptotic Enumeration of Compacted Binary Trees Asymptotic Enumeration of Compacted Binary Trees Antoine Genitrini Bernhard Gittenberger Manuel Kauers Michael Wallner March 30, 207 Abstract A compacted tree is a graph created from a binary tree such

More information

Insert Sorted List Insert as the Last element (the First element?) Delete Chaining. 2 Slide courtesy of Dr. Sang-Eon Park

Insert Sorted List Insert as the Last element (the First element?) Delete Chaining. 2 Slide courtesy of Dr. Sang-Eon Park 1617 Preview Data Structure Review COSC COSC Data Structure Review Linked Lists Stacks Queues Linked Lists Singly Linked List Doubly Linked List Typical Functions s Hash Functions Collision Resolution

More information

Convergence rates in generalized Zeckendorf decomposition problems

Convergence rates in generalized Zeckendorf decomposition problems Convergence rates in generalized Zeckendorf decomposition problems Ray Li (ryli@andrew.cmu.edu) 1 Steven J Miller (sjm1@williams.edu) 2 Zhao Pan (zhaop@andrew.cmu.edu) 3 Huanzhong Xu (huanzhox@andrew.cmu.edu)

More information

Partial Quicksort. Conrado Martínez

Partial 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 information

RANDOM CUTTING AND RECORDS IN DETERMINISTIC AND RANDOM TREES

RANDOM CUTTING AND RECORDS IN DETERMINISTIC AND RANDOM TREES RANDOM CUTTING AND RECORDS IN DETERMINISTIC AND RANDOM TREES SVANTE JANSON Abstract. We study random cutting down of a rooted tree and show that the number of cuts is equal (in distribution) to the number

More information

Hashing. Why Hashing? Applications of Hashing

Hashing. Why Hashing? Applications of Hashing 12 Hashing Why Hashing? Hashing A Search algorithm is fast enough if its time performance is O(log 2 n) For 1 5 elements, it requires approx 17 operations But, such speed may not be applicable in real-world

More information

Lecture: Analysis of Algorithms (CS )

Lecture: Analysis of Algorithms (CS ) Lecture: Analysis of Algorithms (CS483-001) Amarda Shehu Spring 2017 1 Outline of Today s Class 2 Choosing Hash Functions Universal Universality Theorem Constructing a Set of Universal Hash Functions Perfect

More information

On trees with a given degree sequence

On trees with a given degree sequence On trees with a given degree sequence Osvaldo ANGTUNCIO-HERNÁNDEZ and Gerónimo URIBE BRAVO Instituto de Matemáticas Universidad Nacional Autónoma de México 3rd Bath-UNAM-CIMAT workshop 16-05-2016 Our discrete

More information

k-protected VERTICES IN BINARY SEARCH TREES

k-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 information

A Survey of Parking Functions

A Survey of Parking Functions A Survey of Parking Functions Richard P. Stanley M.I.T. Parking functions... n 2 1... a a a 1 2 n Car C i prefers space a i. If a i is occupied, then C i takes the next available space. We call (a 1,...,a

More information

Ascending Runs in Cayley Trees and Mappings

Ascending 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 information

Brownian surfaces. Grégory Miermont based on ongoing joint work with Jérémie Bettinelli. UMPA, École Normale Supérieure de Lyon

Brownian surfaces. Grégory Miermont based on ongoing joint work with Jérémie Bettinelli. UMPA, École Normale Supérieure de Lyon Brownian surfaces Grégory Miermont based on ongoing joint work with Jérémie Bettinelli UMPA, École Normale Supérieure de Lyon Clay Mathematics Institute conference Advances in Probability Oxford, Sept

More information

Algorithms for Combinatorial Systems: Between Analytic Combinatorics and Species Theory.

Algorithms for Combinatorial Systems: Between Analytic Combinatorics and Species Theory. Algorithms for Combinatorial Systems: Between Analytic Combinatorics and Species Theory. Carine Pivoteau - LIGM - december 2011 joint work with Bruno Salvy and Michèle Soria Algorithms for Combinatorial

More information

ROSENA R.X. DU AND HELMUT PRODINGER. Dedicated to Philippe Flajolet ( )

ROSENA 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 information

Application: Bucket Sort

Application: Bucket Sort 5.2.2. Application: Bucket Sort Bucket sort breaks the log) lower bound for standard comparison-based sorting, under certain assumptions on the input We want to sort a set of =2 integers chosen I+U@R from

More information

Data Structures and Algorithm. Xiaoqing Zheng

Data Structures and Algorithm. Xiaoqing Zheng Data Structures and Algorithm Xiaoqing Zheng zhengxq@fudan.edu.cn Dictionary problem Dictionary T holding n records: x records key[x] Other fields containing satellite data Operations on T: INSERT(T, x)

More information

The expansion of random regular graphs

The expansion of random regular graphs The expansion of random regular graphs David Ellis Introduction Our aim is now to show that for any d 3, almost all d-regular graphs on {1, 2,..., n} have edge-expansion ratio at least c d d (if nd is

More information

1 Mechanistic and generative models of network structure

1 Mechanistic and generative models of network structure 1 Mechanistic and generative models of network structure There are many models of network structure, and these largely can be divided into two classes: mechanistic models and generative or probabilistic

More information

Hashing. Martin Babka. January 12, 2011

Hashing. Martin Babka. January 12, 2011 Hashing Martin Babka January 12, 2011 Hashing Hashing, Universal hashing, Perfect hashing Input data is uniformly distributed. A dynamic set is stored. Universal hashing Randomised algorithm uniform choice

More information

Analytic 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. 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 information

Almost all trees have an even number of independent sets

Almost 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 information

Additive tree functionals with small toll functions and subtrees of random trees

Additive tree functionals with small toll functions and subtrees of random trees Additive tree functionals with small toll functions and subtrees of random trees Stephan Wagner To cite this version: Stephan Wagner. Additive tree functionals with small toll functions and subtrees of

More information

Some reflections on directed lattice paths.

Some reflections on directed lattice paths. AofA 4 Paris (June 6-20, 204) DMTCS proc. BA, 204, 25 36 Some reflections on directed lattice paths. Cyril Banderier and Michael Wallner 2 arxiv:605.0687v [cs.dm] 5 May 206 Laboratoire d Informatique de

More information

Partitions with Distinct Multiplicities of Parts: On An Unsolved Problem Posed By Herbert Wilf

Partitions with Distinct Multiplicities of Parts: On An Unsolved Problem Posed By Herbert Wilf Partitions with Distinct Multiplicities of Parts: On An Unsolved Problem Posed By Herbert Wilf James Allen Fill Department of Applied Mathematics and Statistics The Johns Hopkins University 3400 N. Charles

More information

Lecture Notes 7 Random Processes. Markov Processes Markov Chains. Random Processes

Lecture Notes 7 Random Processes. Markov Processes Markov Chains. Random Processes Lecture Notes 7 Random Processes Definition IID Processes Bernoulli Process Binomial Counting Process Interarrival Time Process Markov Processes Markov Chains Classification of States Steady State Probabilities

More information

Permutations with Inversions

Permutations with Inversions 1 2 3 47 6 23 11 Journal of Integer Sequences, Vol. 4 2001, Article 01.2.4 Permutations with Inversions Barbara H. Margolius Cleveland State University Cleveland, Ohio 44115 Email address: b.margolius@csuohio.edu

More information

1 Maintaining a Dictionary

1 Maintaining a Dictionary 15-451/651: Design & Analysis of Algorithms February 1, 2016 Lecture #7: Hashing last changed: January 29, 2016 Hashing is a great practical tool, with an interesting and subtle theory too. In addition

More information

Diagonals: Combinatorics, Asymptotics and Computer Algebra

Diagonals: Combinatorics, Asymptotics and Computer Algebra Diagonals: Combinatorics, Asymptotics and Computer Algebra Bruno Salvy Inria & ENS de Lyon Families of Generating Functions (a n ) 7! A(z) := X n 0 a n z n counts the number of objects of size n captures

More information

THE DEGREE DISTRIBUTION OF RANDOM PLANAR GRAPHS

THE 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 information

Almost giant clusters for percolation on large trees

Almost giant clusters for percolation on large trees for percolation on large trees Institut für Mathematik Universität Zürich Erdős-Rényi random graph model in supercritical regime G n = complete graph with n vertices Bond percolation with parameter p(n)

More information

Random Graphs. 7.1 Introduction

Random Graphs. 7.1 Introduction 7 Random Graphs 7.1 Introduction The theory of random graphs began in the late 1950s with the seminal paper by Erdös and Rényi [?]. In contrast to percolation theory, which emerged from efforts to model

More information

are the q-versions of n, n! and . The falling factorial is (x) k = x(x 1)(x 2)... (x k + 1).

are the q-versions of n, n! and . The falling factorial is (x) k = x(x 1)(x 2)... (x k + 1). Lecture A jacques@ucsd.edu Notation: N, R, Z, F, C naturals, reals, integers, a field, complex numbers. p(n), S n,, b(n), s n, partition numbers, Stirling of the second ind, Bell numbers, Stirling of the

More information

Short cycles in random regular graphs

Short cycles in random regular graphs Short cycles in random regular graphs Brendan D. McKay Department of Computer Science Australian National University Canberra ACT 0200, Australia bdm@cs.anu.ed.au Nicholas C. Wormald and Beata Wysocka

More information

1 Solving Linear Recurrences

1 Solving Linear Recurrences Solving Linear Recurrences Suppose we have a sequence a n defined by the formula The first couple terms of the sequence are a n = 3a n +, a 0 =, 4, 3, 40, 2,... Can you find a general closed for expression

More information

Parking cars after a trailer

Parking cars after a trailer AUSTRALASIAN JOURNAL OF COMBINATORICS Volume 70(3) (2018), Pages 402 406 Parking cars after a trailer Richard Ehrenborg Alex Happ Department of Mathematics University of Kentucky Lexington, KY 40506 U.S.A.

More information

Randomized Load Balancing:The Power of 2 Choices

Randomized Load Balancing:The Power of 2 Choices Randomized Load Balancing: The Power of 2 Choices June 3, 2010 Balls and Bins Problem We have m balls that are thrown into n bins, the location of each ball chosen independently and uniformly at random

More information

GENERALIZED STIRLING PERMUTATIONS, FAMILIES OF INCREASING TREES AND URN MODELS

GENERALIZED 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 information

arxiv: v1 [math.pr] 21 Mar 2014

arxiv: v1 [math.pr] 21 Mar 2014 Asymptotic distribution of two-protected nodes in ternary search trees Cecilia Holmgren Svante Janson March 2, 24 arxiv:4.557v [math.pr] 2 Mar 24 Abstract We study protected nodes in m-ary search trees,

More information

n! (k 1)!(n k)! = F (X) U(0, 1). (x, y) = n(n 1) ( F (y) F (x) ) n 2

n! (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 information

ANATOMY OF THE GIANT COMPONENT: THE STRICTLY SUPERCRITICAL REGIME

ANATOMY OF THE GIANT COMPONENT: THE STRICTLY SUPERCRITICAL REGIME ANATOMY OF THE GIANT COMPONENT: THE STRICTLY SUPERCRITICAL REGIME JIAN DING, EYAL LUBETZKY AND YUVAL PERES Abstract. In a recent work of the authors and Kim, we derived a complete description of the largest

More information

Cookie Monster Meets the Fibonacci Numbers. Mmmmmm Theorems!

Cookie Monster Meets the Fibonacci Numbers. Mmmmmm Theorems! Cookie Monster Meets the Fibonacci Numbers. Mmmmmm Theorems! Steven J. Miller (MC 96) http://www.williams.edu/mathematics/sjmiller/public_html Yale University, April 14, 2014 Introduction Goals of the

More information

WXML Final Report: Chinese Restaurant Process

WXML Final Report: Chinese Restaurant Process WXML Final Report: Chinese Restaurant Process Dr. Noah Forman, Gerandy Brito, Alex Forney, Yiruey Chou, Chengning Li Spring 2017 1 Introduction The Chinese Restaurant Process (CRP) generates random partitions

More information

On the distribution of random walk hitting times in random trees.

On the distribution of random walk hitting times in random trees. On the distribution of random walk hitting times in random trees. Joubert Oosthuizen and Stephan Wagner Abstract The hitting time H xy between two vertices x and y of a graph is the average time that the

More information

What you learned in Math 28. Rosa C. Orellana

What you learned in Math 28. Rosa C. Orellana What you learned in Math 28 Rosa C. Orellana Chapter 1 - Basic Counting Techniques Sum Principle If we have a partition of a finite set S, then the size of S is the sum of the sizes of the blocks of the

More information

CS 125 Section #12 (More) Probability and Randomized Algorithms 11/24/14. For random numbers X which only take on nonnegative integer values, E(X) =

CS 125 Section #12 (More) Probability and Randomized Algorithms 11/24/14. For random numbers X which only take on nonnegative integer values, E(X) = CS 125 Section #12 (More) Probability and Randomized Algorithms 11/24/14 1 Probability First, recall a couple useful facts from last time about probability: Linearity of expectation: E(aX + by ) = ae(x)

More information

The Contour Process of Crump-Mode-Jagers Branching Processes

The Contour Process of Crump-Mode-Jagers Branching Processes The Contour Process of Crump-Mode-Jagers Branching Processes Emmanuel Schertzer (LPMA Paris 6), with Florian Simatos (ISAE Toulouse) June 24, 2015 Crump-Mode-Jagers trees Crump Mode Jagers (CMJ) branching

More information

INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET AUTOMATIQUE. The Average Case Analysis of. Multivariate Asymptotics and N 3162.

INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET AUTOMATIQUE. The Average Case Analysis of. Multivariate Asymptotics and N 3162. INSTITUT NATIONAL DE RECHERCHE EN INFORMATIQUE ET AUTOMATIQUE The Average Case Analysis of Algorithms: Multivariate Asymptotics and Limit Distributions Philippe FLAJOLET, Robert SEDGEWICK 023456789 N 362

More information

IEOR 8100: Topics in OR: Asymptotic Methods in Queueing Theory. Fall 2009, Professor Whitt. Class Lecture Notes: Wednesday, September 9.

IEOR 8100: Topics in OR: Asymptotic Methods in Queueing Theory. Fall 2009, Professor Whitt. Class Lecture Notes: Wednesday, September 9. IEOR 8100: Topics in OR: Asymptotic Methods in Queueing Theory Fall 2009, Professor Whitt Class Lecture Notes: Wednesday, September 9. Heavy-Traffic Limits for the GI/G/1 Queue 1. The GI/G/1 Queue We will

More information

Massachusetts Institute of Technology

Massachusetts Institute of Technology .203J/6.28J/3.665J/5.073J/6.76J/ESD.26J Quiz Solutions (a)(i) Without loss of generality we can pin down X at any fixed point. X 2 is still uniformly distributed over the square. Assuming that the police

More information

Homework 6: Solutions Sid Banerjee Problem 1: (The Flajolet-Martin Counter) ORIE 4520: Stochastics at Scale Fall 2015

Homework 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 information

The Moments of the Profile in Random Binary Digital Trees

The 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 information

On 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) 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 information

Conjecture 4. f 1 (n) = n

Conjecture 4. f 1 (n) = n 26 International Conference on Computational Science and Computational Intelligence Computation in the Enumeration of Integer Matrices Shanzhen Gao, Keh-Hsun Chen Department of Computer Science, College

More information