On the distribution of the longest run in number partitions

We consider the distribution of the longest run of equal elements in number partitions (equivalently, the distribution of the largest gap between subsequent elements); in a recent paper, Mutafchiev proved that the distribution of this random variable (appropriately rescaled) converges weakly. The corresponding distribution function is closely related to the generating function for number partitions. In this paper, this problem is considered in more detail—we study the behavior at the tails (especially the case that the longest run is comparatively small) and extend the asymptotics for the distribution function to the entire interval of possible values. Additionally, we prove a local limit theorem within a suitable region, i.e. when the longest run attains its typical order n ^1/2, and we observe another phase transition that occurs when the largest gap is of order n ^1/4: there, the conditional probability that the longest run has length d, given that it is ≤d, jumps from 1 to 0. Asymptotics for the mean and variance follow immediately from our considerations.     

Regularity characterization of asymptotically probability equivalent sequences

The notion of asymptotically equivalent sequences was presented by Pobyvanets in 1980. Using this definition, he presented Silverman–Toeplitz-type matrix conditions that ensure that a summability matrix preserves asymptotic equivalency. This work begins with an extension of Pobyvanets’ definition of convergence in probability. This definition is also used to present Silverman–Toeplitz-type conditions for ensuring that a summability matrix preserves asymptotic probability equivalence. In addition, we shall also present a Marouf-type characterization of such a sequence space.     

On long runs of heads and tails II

In a string ofn independent coin tosses we consider the difference between the lengths of the longest blocks of consecutive heads resp. tails. A complete characterization of the a.s. limit properties of this quantity is proved.     

Coding into Ramsey sets

In [6] W. T. Gowers formulated and proved a Ramsey-type result which lies at the heart of his famous dichotomy for Banach spaces. He defines the notion of weakly Ramsey set of block sequences of an infinite dimensional Banach space and shows that every analytic set of block sequences is weakly Ramsey. We show here that Gowers’ result follows quite directly from the fact that all G_δ sets are weakly Ramsey, if the Banach space does not contain c₀, and from the fact that all F_σδ sets are weakly Ramsey, in the case of an arbitrary Banach space. We also show that every result obtained by the application of Gowers’ theorem to an analytic set can also be obtained by applying the Theorem to a F_σδ set (or to a G_δ set if the space does not contain c₀). This fact explains why the only known applications of this technique are based on very low-ranked Borel sets (open, closed, F_σ, or G_δ).     

A note on runs of geometrically distributed random variables

Recently, Grabner et al. [Combinatorics of geometrically distributed random variables: run statistics, Theoret. Comput. Sci. 297 (2003) 261-270] and Louchard and Prodinger [Ascending runs of sequences of geometrically distributed random variables: a probabilistic analysis, Theoret. Comput. Sci. 304 (2003) 59-86] considered the run statistics of geometrically distributed independent random variables. They investigated the asymptotic properties of the number of runs and the longest run using the corresponding probability generating functions and a Markov chain approach. In this note, we reconsider the asymptotic properties of such statistics using another approach. Our approach of finding the asymptotic distributions is based on the construction of runs in a sequence of m-dependent random variables. This approach enables us to find the asymptotic distributions of many run statistics via the theorems established for m-dependent sequence of random variables. We also provide the asymptotic distribution of the total number of non-decreasing runs and the longest non-decreasing run.     

An extreme value theory for long head runs

Summary For an infinite sequence of independent coin tosses withP(Heads)=p(0,1), the longest run of consecutive heads in the firstn tosses is a natural object of study. We show that the probabilistic behavior of the length of the longest pure head run is closely approximated by that of the greatest integer function of the maximum ofn(1-p) i.i.d. exponential random variables. These results are extended to the case of the longest head run interrupted byk tails. The mean length of this run is shown to be log(n)+klog(n)+(k+1)log(1–p)–log(k!)+k+/–1/2+ r₁(n)+ o(1) where log=log_1/p , =0.577 ... is the Euler-Mascheroni constant, =ln(1/p), andr ₁(n) is small. The variance is ₂/6²+1/12 +r ₂(n)+ o(1), wherer ₂(n) is again small. Upper and lower class results for these run lengths are also obtained and extensions discussed.This work was supported by a grant from the System Development Foundation     

Generalized Greatest Common Divisors, Divisibility Sequences, and Vojta’s Conjecture for Blowups

We apply Vojta’s conjecture to blowups and deduce a number of deep statements regarding (generalized) greatest common divisors on varieties, in particular on projective space and on abelian varieties. Special cases of these statements generalize earlier results and conjectures. We also discuss the relationship between generalized greatest common divisors and the divisibility sequences attached to algebraic groups, and we apply Vojta’s conjecture to obtain a strong bound on the divisibility sequences attached to abelian varieties of dimension at least two.     

Generalized Greatest Common Divisors,Divisibility Sequences,and Vojta’s Conjecture for Blowups

We apply Vojta’s conjecture to blowups and deduce a number of deep statements regarding (generalized) greatest common divisors on varieties, in particular on projective space and on abelian varieties. Special cases of these statements generalize earlier results and conjectures. We also discuss the relationship between generalized greatest common divisors and the divisibility sequences attached to algebraic groups, and we apply Vojta’s conjecture to obtain a strong bound on the divisibility sequences attached to abelian varieties of dimension at least two.     

Long repetitive patterns in random sequences

Summary Appearances of long repetitive sequences such as 00...0 or 1010...101 in random sequences are studied. The expected length of the longest repetitive run of any specified type in a random binary sequence of length n is shown to tend to the binary logarithm of n plus a periodic function of log n. Necessary and sufficient conditions are derived to ensure that with probability 1 an infinite random sequence should contain repetitive runs of specified lengths in given initial segments. Finally, the number of long repetitive runs of a specified kind that occur in a random sequence is studied. These results are derived from simple expressions for the generating functions for the probabilities of occurrences of various repetitive runs. These generating functions are rational, and lead to sharp asymptotic estimates for the probabilities.     

Determining the form of the mean of a stochastic process

Let μ be the mean function of an observable stochastic process whose sample paths fall in some Banach space with a basis and assume μ is also in this space. A procedure like Cover's (Ann. Statist.1, 862–871, 1973) is given which has the property that if the last nonzero coordinate of μ is the mth then with probability one this is discovered after at most a finite number of erros. If μ has an infinite number of nonzero coordinates, then with probability one this is discovered after at most a finite number of errors except for a set of μ of prior probability zero.