Poissonized Model in Combinatorics & Probability

• Poissonized model is a technique that replaces fixed discrete indices with a Poisson random variable or process to create additivity and independence.
• It enables precise asymptotic analysis in fields like random graphs, partitions, and stochastic geometry by allowing tractable de-Poissonization methods.
• Its applications range from longest increasing subsequence problems to kernel determinations in random partitions and learning-theoretic Markov algorithms.

Searching arXiv for recent and foundational papers on Poissonized models across combinatorics, probability, and stochastic processes. A Poissonized model is a reformulation in which a fixed discrete size, count, or iteration index is replaced by a Poisson random variable, a Poisson point process, or a Poisson clock. In the cited literature, this device appears in asymptotic analysis of longest increasing subsequences, random graphs, random partitions, coupon collection, stochastic geometry, branching systems, and learning-theoretic Markov algorithms. Its recurring role is to convert fixed-size dependence into structures with exact additivity, independent increments, determinantal kernels, or continuous-time semigroups, after which one may either remain in the Poissonized setting or recover the fixed-size model by de-Poissonization (Bornemann, 2023, Curien, 2022, Dupuis et al., 11 Feb 2025).

1. Core constructions

Across the literature, Poissonization takes several standard forms. In analytic combinatorics and asymptotic representation theory, one introduces a Poisson transform of a fixed-$n$ quantity. For the longest increasing subsequence (LIS) distribution, if $L_n(\cdot)$ is the CDF of the LIS length in a random $n$-permutation, its Poisson generating function is

$$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$

and Cauchy’s formula recovers the fixed-$n$ probability from its Poisson transform (Bornemann, 2023).

In probabilistic combinatorics, Poissonization often means randomizing the underlying population size. In the Poissonized Erdős–Rényi model, one fixes $\alpha>0$, lets $N\sim\mathrm{Pois}(\alpha)$, and then forms $G(N,p)$ on a core of $N$ vertices, together with an infinite stack of extra vertices attached only to the core (Curien, 2022). In stochastic geometry, one replaces $n$ i.i.d. points by a Poisson point process. For boundary polytopes, $L_n(\cdot)$0 is a Poisson point process on $L_n(\cdot)$1 with intensity measure $L_n(\cdot)$2, and the Poissonized polytope is $L_n(\cdot)$3 (Reitzner et al., 24 Sep 2025). In nonhomogeneous geometric CLTs, the Poissonized version of the binomial process is $L_n(\cdot)$4 with $L_n(\cdot)$5 independent of the $L_n(\cdot)$6 (Trinh, 2018).

A third construction inserts a Poisson clock into a discrete-time process. For a time-homogeneous Markov chain $L_n(\cdot)$7, the Poissonized process is

$L_n(\cdot)$8

where $L_n(\cdot)$9 is an independent unit-rate Poisson clock. Its semigroup is

$n$0

and its infinitesimal generator is $n$1 (Dupuis et al., 11 Feb 2025).

Construction	Representative formula	Typical consequence
Poisson transform	$n$2	asymptotic expansion, de-Poissonization
Random size	$n$3	exact Markov or independence structure
Poisson point process	$n$4	stabilization, variance asymptotics, CLT
Poisson clock	$n$5	continuous-time jump process with generator $n$6

This suggests that “Poissonized model” is less a single model class than a general analytic and probabilistic device. The unifying feature is that Poisson randomness is introduced exactly where it yields factorization or a tractable semigroup.

2. Poissonization as an asymptotic analytic device

A central use of Poissonization is to derive asymptotic expansions in a setting where the Poissonized object is more accessible than the fixed-size one. For LIS, one studies $n$7 in the edge scaling

$n$8

and obtains the expansion

$n$9

where $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$0 is the GUE Tracy–Widom CDF and $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$1 are explicit linear combinations of derivatives of $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$2 with polynomial coefficients in $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$3 (Bornemann, 2023). The first correction terms are written explicitly in that work, and higher $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$4 remain explicit linear combinations of $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$5.

The analytic gain comes from an exact identity between the Poissonized LIS distribution and the hard-edge gap probability of LUE,

$$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$6

followed by a hard-to-soft edge transition. Uniform Olver-type expansions for $$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$7 in the transition region and a kernel expansion from the Bessel determinant to the Airy kernel yield the Poissonized asymptotic expansion (Bornemann, 2023).

Recovery of the fixed-$$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$8 model is a separate step. The same paper replaces Johansson’s monotonicity-based de-Poissonization by analytic de-Poissonization of Jacquet and Szpankowski, using the contour integral

$$P_\lambda(k)=e^{-\lambda}\sum_{n=0}^{\infty}(L_n\le k)\frac{\lambda^n}{n!},$$9

This requires a tameness hypothesis: for $n$0 in the asymptotic window $n$1, the entire function $n$2 should have no zeros too close to the positive real axis and no zeros of size $n$3 in a sector $n$4 (Bornemann, 2023).

A parallel pattern appears in stochastic geometry. Poissonization gives exact spatial independence between disjoint blocks, after which one uses block sums, the Poisson Poincaré inequality, stabilization, and then classical de-Poissonization arguments to recover a fixed-$n$5 CLT for the binomial process (Trinh, 2018). In the Poissonized Erdős–Rényi setting, a standard “sandwich”/depoissonization argument transfers the connectivity limit from $n$6 to $n$7 (Curien, 2022).

A common misconception is that Poissonization by itself settles the original fixed-size problem. The cited works show the opposite: the Poissonized model is often the analytically tractable object, but fixed-size conclusions may depend on contour methods, sandwiching, or extra growth and moment conditions (Bornemann, 2023, Trinh, 2018).

3. Determinantal and partition-theoretic Poissonized models

Poissonization is especially prominent in random partition theory because it exposes determinantal structure. For partitions $n$8, the Poissonized Plancherel measure is

$n$9

and under the map

$\alpha>0$0

it becomes a determinantal point process with the discrete Bessel kernel $\alpha>0$1 (Lazag, 2019). In this setting, Christoffel deformations remain determinantal, and the reduced Palm measure at points $\alpha>0$2 coincides with the corresponding Christoffel deformation (Lazag, 2019).

The periodic Schur process provides a second Poissonized partition model. The ordinary Poissonized Plancherel measure arises as the $\alpha>0$3 specialization of the Schur process with exponential specialization $\alpha>0$4, while the cylindric deformation introduces an extra parameter $\alpha>0$5 and produces a determinantal point process after passing to a grand canonical ensemble (Betea et al., 2018). In the edge crossover regime

$\alpha>0$6

the rescaled kernel converges to the finite-temperature Airy kernel $\alpha>0$7, and the extreme-value statistics interpolate between Tracy–Widom GUE and Gumbel (Betea et al., 2018).

The Poissonized Robinson–Schensted process gives a dynamical version of the same theme. Starting from a rate-one Poisson point process $\alpha>0$8 in $\alpha>0$9, the decorated Robinson–Schensted correspondence produces a random pair of decorated Young tableaux and hence a Young-diagram process $N\sim\mathrm{Pois}(\alpha)$0, $N\sim\mathrm{Pois}(\alpha)$1 (1404.02823). Its finite-dimensional distributions form a Schur process, its fixed-time marginal is Poissonized Plancherel with parameter $N\sim\mathrm{Pois}(\alpha)$2, and the associated line ensemble is equivalent to non-intersecting Poisson arches (1404.02823).

Recent asymptotic work shows that Poissonized Plancherel models also support multiplicative observables with nontrivial phase structure. For the expectation

$N\sim\mathrm{Pois}(\alpha)$3

the large-$N\sim\mathrm{Pois}(\alpha)$4 asymptotics of $N\sim\mathrm{Pois}(\alpha)$5 produce a rate function $N\sim\mathrm{Pois}(\alpha)$6, explicit oscillatory terms in elliptic theta functions, and two third-order phase transitions of different nature (Cafasso et al., 8 Jan 2026).

4. Random graphs, coupon collection, and tree dynamics

In random graph theory, Poissonization can transform exploration processes into objects with independent increments. In $N\sim\mathrm{Pois}(\alpha)$7, if $N\sim\mathrm{Pois}(\alpha)$8 is the top stack vertex and $N\sim\mathrm{Pois}(\alpha)$9 is its number of neighbors in the core, then $G(N,p)$0, and after removing $G(N,p)$1 and replacing it by its $G(N,p)$2 core-neighbors in the stack, the remaining graph is again $G(N,p)$3, independent of $G(N,p)$4 (Curien, 2022). Consequently, the Lukasiewicz exploration has independent Poisson increments

$G(N,p)$5

and

$G(N,p)$6

for a standard rate-1 Poisson process $G(N,p)$7 (Curien, 2022). This yields short proofs of the giant-component transition at $G(N,p)$8, the connectivity threshold $G(N,p)$9, a CLT for the giant component, and Aldous’s critical scaling window (Curien, 2022).

In the double Dixie cup problem, Poissonization replaces discrete coupon draws by $N$0 independent Poisson processes of rates $N$1. The completion time for obtaining $N$2 copies of every coupon is then

$N$3

where each $N$4 is Erlang$N$5 and the $N$6 are independent across $N$7 (Long, 28 Apr 2026). The product-form CDF

$N$8

drives a strict variance-extremality theorem: for every fixed $N$9 and $n$0, the variance of $n$1 is uniquely minimized by the uniform coupon vector $n$2 (Long, 28 Apr 2026).

The Poissonized Aldous chain similarly replaces discrete leaf moves on binary trees by independent exponential clocks: each existing leaf has a death clock of rate $n$3, and each existing interior edge has a birth clock of rate $n$4 (Pal, 2011). Under rescaling of time by $n$5, edge lengths by $n$6, and leaf masses by $n$7, the finite Poissonized chain converges to a continuum-tree process encoded by a contour Lévy process and a Poisson additive point process of ages (Pal, 2011).

These examples show that Poissonization does not merely smooth counts. It can expose exact branching, exploration, or extremal structure that is obscure in the original discrete model.

5. Point processes, stabilization, and measure-valued limits

In stochastic geometry, the Poissonized model is frequently the natural rather than auxiliary object. For random boundary polytopes, $n$8 is a Poisson point process on $n$9 with intensity $L_n(\cdot)$00, and

$L_n(\cdot)$01

admits a score decomposition because $L_n(\cdot)$02 is almost surely simplicial in the smooth case (Reitzner et al., 24 Sep 2025). The radius of stabilization $L_n(\cdot)$03 satisfies an exponential tail bound,

$L_n(\cdot)$04

which feeds second-order Poincaré inequalities and quantitative normal approximation. The result is order-$L_n(\cdot)$05 expectation and variance asymptotics together with a CLT and an optimal Berry–Esseen bound $L_n(\cdot)$06 for $L_n(\cdot)$07 when $L_n(\cdot)$08 and $L_n(\cdot)$09 (Reitzner et al., 24 Sep 2025).

Kurtz and Rodrigues use a Poissonized particle representation for branching Markov processes and their measure-valued limits. In the finite-$L_n(\cdot)$10 model, particles carry locations and levels, birth occurs at rate $L_n(\cdot)$11, and death occurs when a level reaches $L_n(\cdot)$12 (Kurtz et al., 2011). At fixed time $L_n(\cdot)$13, conditioned on the spatial configuration, levels are independent and uniformly distributed on $L_n(\cdot)$14. In the limit $L_n(\cdot)$15, conditioned on the random measure $L_n(\cdot)$16, the point measure

$L_n(\cdot)$17

is a Poisson random measure on $L_n(\cdot)$18 with mean intensity $L_n(\cdot)$19 (Kurtz et al., 2011). This yields the Laplace functional of the associated Dawson–Watanabe superprocess.

More generally, CLTs for stabilizing functionals on Poisson point processes exploit the add-one cost

$L_n(\cdot)$20

together with the Poisson Poincaré inequality

$L_n(\cdot)$21

The Poissonized framework then supports homogeneous and nonhomogeneous CLTs before de-Poissonization transfers them to binomial input (Trinh, 2018).

6. Statistical and algorithmic uses, scope, and limitations

Outside classical probability, Poissonized models appear as inferential or algorithmic surrogates. In POI-SIMEX for tissue histology, the observed biomarker count in a tissue core satisfies

$L_n(\cdot)$22

so the measurement error $L_n(\cdot)$23 is non-Gaussian and heteroscedastic with

$L_n(\cdot)$24

The SIMEX correction then adds simulated noise at levels $L_n(\cdot)$25 and extrapolates to $L_n(\cdot)$26, yielding a strongly consistent estimator under the conditional Poisson surrogate model in linear regression (Yang et al., 2024).

In learning theory, Poissonization converts a discrete-time Markov learning algorithm into a continuous-time jump process with generator $L_n(\cdot)$27, restoring a closed-form entropy flow. If $L_n(\cdot)$28 and $L_n(\cdot)$29 are the prior and posterior densities, then

$L_n(\cdot)$30

and the relative entropy satisfies

$L_n(\cdot)$31

This directly yields PAC-Bayesian generalization bounds and links to modified logarithmic Sobolev inequalities (Dupuis et al., 11 Feb 2025).

The range of these examples clarifies two limitations. First, Poissonization is not synonymous with “Poisson model” in the narrow sense. For example, a Poisson degree distribution in the Poisson-network SIR model yields the identity $L_n(\cdot)$32 and exact edge-based closure, but this is a structural Poisson assumption rather than a de-Poissonization scheme (Wairimu et al., 2024). Second, Poissonization is not automatically exact for the original non-Poissonized problem. LIS asymptotics require a tameness hypothesis for analytic de-Poissonization, geometric CLTs require extra moment-growth conditions, and fixed-size graph results use sandwiching arguments (Bornemann, 2023, Trinh, 2018, Curien, 2022).

Taken together, these works present the Poissonized model as a versatile probabilistic technology. It may serve as a generating transform, a random-size embedding, a Poisson point-process replacement, or a Poisson-clock interpolation. The main mathematical payoff is that Poissonization frequently converts a difficult discrete object into one governed by independent increments, product-form distributions, determinantal kernels, stabilization estimates, or a simple continuous-time generator, while preserving a route back to the original fixed-size model when de-Poissonization is available (Bornemann, 2023, Lazag, 2019, Dupuis et al., 11 Feb 2025).