Abstract:Assume that $2n$ balls are thrown independently and uniformly at random into $n$ bins. We consider the unlikely event $E$ that every bin receives at least one ball, showing that $\Pr[E] = \Theta(b^n)$ where $b \approx 0.836$. Note that, due to correlations, $b$ is not simply the probability that any single bin receives at least one ball. More generally, we consider the event that throwing $\alpha n$ balls into $n$ bins results in at least $d$ balls in each bin.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the probability that each bucket receives at least a certain number of balls when balls are randomly distributed into several buckets under given conditions. Specifically, assume that $2n$ balls are independently and uniformly randomly thrown into $n$ buckets. Consider the event $E$ that each bucket has at least one ball. The author shows the relationship between the probability $Pr[E]$ of this event and a certain constant $b\approx0.836$, that is, $Pr[E]=\Theta(b^{n})$. It should be noted here that due to the correlation between the buckets, $b$ is not the probability that a single bucket receives at least one ball. More generally, the paper also explores the probability that each bucket receives at least $d$ balls when $\alpha n$ balls are thrown into $n$ buckets. Here, $\alpha$ and $d$ are parameters, and $\alpha\geq d$. By introducing a special distribution $\Phi(\alpha, d)$, which is a Poisson distribution truncated to values $\geq d$ and adjusted so that its expectation is $\alpha$, the paper obtains the following results: 1. When $\alpha$ and $d$ are constants and $\alpha > d$, $Pr[E]=\Theta(b^{n})$, where $b = \frac{\alpha^{\alpha}\zeta}{e^{\alpha}\lambda^{\alpha - d}}$. 2. When $\alpha = d$ (not necessarily a constant), $Pr[E]=\Theta(b^{n}\sqrt{\frac{d}{n}})$, where $b=\frac{d^{d}}{e^{d}d!}$. The main contribution of the paper lies in providing an accurate mathematical framework for calculating the above - mentioned probabilities under different conditions, which is of great significance for understanding problems in computer science fields such as random hash functions and minimum perfect hash functions. In addition, the paper also provides specific calculation methods and some numerical results, which are helpful for reference and verification in practical applications.

The Probability to Hit Every Bin with a Linear Number of Balls

ON THE FLUCTUATION IN THE RANDOM ASSIGNMENT PROBLEM

Expected number of uniformly distributed balls in a most loaded bin using placement with simple linear functions

Avoiding defeat in a balls-in-bins process with feedback

Load Balancing with Dynamic Set of Balls and Bins

Laws of Large Numbers for Weighted Sums of Random Variables and Random Elements

The power of thinning in balanced allocation

An asymptotically optimal algorithm for generating bin cardinalities

Load balancing under $d$-thinning

A Novel Approach to Probability

Parallel Balanced Allocations: The Heavily Loaded Case

A Generalization of Multiple Choice Balls-into-Bins: Tight Bounds

A Note On the Size of Largest Bins Using Placement With Linear Transformations

Some Approximations to the Binomial Distribution Function

The Limit Points of the Strong Law of Large Numbers under the Sub-Linear Expectations

Balanced Allocations: Caching and Packing, Twinning and Thinning

LIL Behavior for B-valued Strong Mixing Random Variables

A probabilistic approach to Lorentz balls

LIL and the Approximation of Rectangular Sums of B-valued Random Variables when Extreme Terms Are Excluded

The Distribution of a Sum of Independent Binomial Random Variables

Sharper analysis of the random graph $d$-process via a balls-in-bins model