[Paper Review] Poisson and independent process approximation for random combinatorial structures with a given number of components, and near-universal behavior for low rank assemblies

Richard Arratia, Stephen DeSalvo|arXiv (Cornell University)|Jun 15, 2016

Advanced Combinatorial Mathematics36 references3 citations

TL;DR

This paper develops a Poisson process approximation framework for random combinatorial assemblies conditioned on a fixed number of components, enabling precise analysis of component size distributions when the rank $ r = n - k $ is small. It establishes a near-universal limit law: for $ r \sim t\sqrt{n} $, the largest component size converges in probability to 2 or 3, with the number of size-3 components converging to a Poisson distribution, resolving a case previously intractable by saddlepoint methods.

ABSTRACT

We give a general framework for approximations to combinatorial assemblies, especially suitable to the situation where the number $k$ of components is specified, in addition to the overall size $n$. This involves a Poisson process, which, with the appropriate choice of parameter, may be viewed as an extension of saddlepoint approximation. We illustrate the use of this by analyzing the component structure when the rank and size are specified, and the rank, $r := n-k$, is small relative to $n$. There is near-universal behavior, in the sense that apart from cases where the exponential generating function has radius of convergence zero, for $\ell=1,2,\dots$, when $r \asymp n^α$ for fixed $α\in (\frac{\ell}{\ell+1}, \frac{\ell+1}{\ell+2})$, the size $L_1$ of the largest component converges in probabiity to $\ell+2$. Further, when $r \sim t\, n^{\ell/(\ell+1)}$ for a positive integer $\ell$, and $t \in (0,\infty)$, $\mathbb{P}\,(L_1 \in \{\ell+1,\ell+2\}) o 1$, with the choice governed by a Poisson limit distribution for the number of components of size $\ell+2$. This was previously observed, for the case $\ell=1$ and the special cases of permutations and set partitions, using Chen-Stein approximations for the indicators of attacks and alignments, when rooks are placed randomly on a triangular board. The case $\ell=1$ is especially delicate, and was not handled by previous saddlepoint approximations.

Motivation & Objective

• To develop a probabilistic framework for approximating random combinatorial assemblies conditioned on a fixed number of components.
• To extend saddlepoint approximation techniques to cases where traditional Gaussian approximations fail, particularly when the rank $ r = n - k $ is small.
• To establish a universal limit law for low-rank assemblies, showing near-universal behavior in component size distribution across diverse combinatorial families.
• To provide both asymptotic and effective quantitative error bounds for the Poisson approximation in the critical regime $ r \sim t\sqrt{n} $.

Proposed method

• Uses a conditional Poisson process with a parameter chosen to match the component count, extending saddlepoint approximation to combinatorial assemblies.
• Applies Chen-Stein method and binomial approximation techniques to control error bounds in the conditional distribution of component sizes.
• Introduces a two-stage sampling approach: first condition on the number of components $ k $, then use Poisson process to model component size counts.
• Employs generating function analysis and asymptotic bounds on $ \mathbb{P}(N_j = m \mid \sum jN_j = r) $ to derive convergence results.
• Derives explicit error bounds via $ u_M(n,k) $, a function of $ n, k $, and assembly parameters, ensuring quantitative control.
• Analyzes the critical regime $ r \sim t\sqrt{n} $ by splitting the probability space into regions where $ N_1 $ is concentrated and using tail bounds on binomial variables.

Experimental results

Research questions

• RQ1 What is the limiting distribution of the largest component size in a random combinatorial assembly with $ n $ elements and $ k $ components when $ r = n - k \sim t\sqrt{n} $?
• RQ2 How can Poisson process approximation be used to model component size counts in assemblies with fixed component count, especially when saddlepoint methods fail?
• RQ3 What universal behavior emerges in low-rank assemblies as $ r \sim t n^{\alpha} $ for $ \alpha \in (\frac{\ell}{\ell+1}, \frac{\ell+1}{\ell+2}) $?
• RQ4 Can effective, non-asymptotic error bounds be derived for the Poisson approximation in the critical regime $ r \sim t\sqrt{n} $?
• RQ5 Why do permutations and set partitions exhibit the same Poisson limit for component counts despite differing global component distributions?

Key findings

• When $ r = n - k \sim t\sqrt{n} $, the largest component size $ L_1 $ satisfies $ \mathbb{P}(L_1 \in \{2,3\}) \to 1 $, with the number of components of size 3 converging to a Poisson distribution.
• For $ r \sim t n^{\ell/(\
• The paper establishes that the Poisson limit law for component counts in low-rank assemblies is near-universal, applying to all assemblies except those with zero radius of convergence in their exponential generating function.
• The asymptotic error bound for the critical regime is $ O_t(\log^2 n / \sqrt{n}) $, which is sharp and quantitatively controlled.
• An effective error bound $ u_M(n,k) $ is derived in Theorem 3.19, providing a fully explicit, non-asymptotic estimate for the approximation error.
• The case $ \ell = 1 $, corresponding to $ r \sim t\sqrt{n} $, is shown to be analytically delicate and previously intractable by standard saddlepoint methods, which this paper resolves.