Step Recurrence: Theory & Applications

• Step recurrence is a concept in discrete mathematics that captures the return or repetition behavior in systems like orthogonal polynomials, random walks, and dynamical systems.
• The recurrence relations in multiple orthogonal polynomials enable algorithmic computation and explicit structural insights through step-line formulations.
• In stochastic and dynamical settings, step recurrence delineates critical thresholds and phase transitions, enhancing practical understanding of return probabilities and system dynamics.

Step recurrence encompasses a constellation of mathematical phenomena in discrete systems, orthogonal polynomial theory, stochastic processes, and dynamical systems, all involving return, repetition, or structure at the level of single-step or multi-step sequences. Its central manifestations include recurrence relations for special functions (notably multiple orthogonal polynomials), step-by-step recurrence/transience behavior in random and quantum walks, phase transitions in step-reinforced processes, and combinatorial recurrence in dynamical systems.

1. Step-Line Recurrence in Multiple Orthogonal Polynomials

For type II multiple orthogonal polynomials with $r$ measures, the step-line recurrence is a canonical $(r+2)$-term relation near the diagonal of the multi-index lattice. Define $\mathbf{n} = (n_1,\ldots,n_r)$ with $|\mathbf{n}| = n_1 + \cdots + n_r$ and let $P_{\mathbf{n}}(x)$ be the unique (monic) polynomial of degree $|\mathbf{n}|$ satisfying

$$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$

On the "step-line" (the sequence of multi-indices closest to the main diagonal), the system admits the recurrence

$$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$

where $p_{rn+j}(x)$ corresponds to moving $j$ steps off the main diagonal and the $(r+2)$0 are step-line recurrence coefficients with explicit boundary conditions $(r+2)$1 for $(r+2)$2, $(r+2)$3, $(r+2)$4 (Filipuk et al., 2014).

For $(r+2)$5, these coefficients are computed explicitly from the nearest-neighbor recurrences via closed formulas, and the step-line recurrence encodes the full structure of the type II system. The step-line and nearest-neighbor recurrences are algorithmically interrelated, and in classical cases (e.g., through the modified Bessel measures), closed-form step-line coefficients exist.

2. Step Recurrence in Random Walks: Classical, Reinforced, and Quantum

Step recurrence in stochastic processes describes the statistical return to specific states or points after a given number of steps, revealing sharp phase transitions.

Step-Reinforced Random Walks (SRRW): In $(r+2)$6 dimensions, an SRRW with reinforcement parameter $(r+2)$7 (the probability to repeat a past step) exhibits a phase transition between recurrence and transience governed by $(r+2)$8:

• For $(r+2)$9, the walk is recurrent if $\mathbf{n} = (n_1,\ldots,n_r)$0, transient if $\mathbf{n} = (n_1,\ldots,n_r)$1.
• For $\mathbf{n} = (n_1,\ldots,n_r)$2, recurrence holds for $\mathbf{n} = (n_1,\ldots,n_r)$3, transience for $\mathbf{n} = (n_1,\ldots,n_r)$4 (with transience even at $\mathbf{n} = (n_1,\ldots,n_r)$5 under stronger moment assumptions).
• For $\mathbf{n} = (n_1,\ldots,n_r)$6, the walk is always transient, regardless of $\mathbf{n} = (n_1,\ldots,n_r)$7.

The critical threshold at $\mathbf{n} = (n_1,\ldots,n_r)$8 separates step recurrence from divergence, controlled analytically via Lyapunov functions and martingale arguments. At criticality, sharp asymptotic scaling laws for the occupation time and empirical occupation measure can be established (Qin, 2024).

Horizontal-Vertical (H–V) Walks on $\mathbf{n} = (n_1,\ldots,n_r)$9: With randomized site labels determining the direction of each step and a label-flipping probability $|\mathbf{n}| = n_1 + \cdots + n_r$0, recurrence for a single walker is guaranteed for $|\mathbf{n}| = n_1 + \cdots + n_r$1; for multiple walkers, the threshold depends on the number of walkers and $|\mathbf{n}| = n_1 + \cdots + n_r$2. This establishes nontrivial step recurrence criteria beyond the classical setting (Chan, 2020).

Rademacher Random Walks with Variable Step Sizes: For walks with i.i.d. $|\mathbf{n}| = n_1 + \cdots + n_r$3 increments scaled by a deterministic sequence $|\mathbf{n}| = n_1 + \cdots + n_r$4, the dichotomy between recurrence and transience is controlled by the growth rate of $|\mathbf{n}| = n_1 + \cdots + n_r$5:

• Bounded $|\mathbf{n}| = n_1 + \cdots + n_r$6 yields (weak) recurrence.
• For $|\mathbf{n}| = n_1 + \cdots + n_r$7, recurrence holds if $|\mathbf{n}| = n_1 + \cdots + n_r$8 and transience if $|\mathbf{n}| = n_1 + \cdots + n_r$9—a sharp transition point.
• Strong anti-concentration estimates and block constructions underpin this step-wise recurrence classification (Bhattacharya et al., 28 Oct 2025).

Discrete-Time Quantum Stochastic Walks (DTQSWs): Quantum walks augmented by classical randomness (parameter $P_{\mathbf{n}}(x)$0) support step recurrence phenomena not present in pure quantum or pure classical walks. Notably, for coin angles above a critical $P_{\mathbf{n}}(x)$1, a small amount of classical randomness suppresses the overall return probability below its purely unitary quantum value before it rises again for larger $P_{\mathbf{n}}(x)$2. The regime and functional form of step recurrence are encoded in the first-return generating function, computed using Fourier-integral techniques (Stefanak et al., 15 Jan 2025).

Explicit Formulas for Return Probabilities: In one-dimensional biased random walks, the survival probability (no return to the origin after $P_{\mathbf{n}}(x)$3 steps) has closed-form expressions as a function of bias parameter $P_{\mathbf{n}}(x)$4 and $P_{\mathbf{n}}(x)$5, exact for all $P_{\mathbf{n}}(x)$6. For large $P_{\mathbf{n}}(x)$7, the survival probability approaches $P_{\mathbf{n}}(x)$8, recovering Pólya's law, but finite-$P_{\mathbf{n}}(x)$9 corrections and last-return statistics (including the loss of the endpoint peak above a critical bias) are analytically accessible (Mookerjee et al., 2024).

3. Structural and Computational Aspects of Step-Line Recurrences

In the algebraic theory of (mixed) multiple orthogonality, the step-line recurrence leads to the construction of a banded, generally non-normal recurrence matrix $|\mathbf{n}|$0:

$|\mathbf{n}|$1

where $|\mathbf{n}|$2 has $|\mathbf{n}|$3 subdiagonals and $|\mathbf{n}|$4 superdiagonals for parameters $|\mathbf{n}|$5. $|\mathbf{n}|$6 can be explicitly factorized into bidiagonal (elementary Christoffel/Darboux) operators using the $|\mathbf{n}|$7 factorization of the block moment matrix, or alternatively by upper-lower duality. The coefficients of each band are given in terms of ratios of leading coefficients of Christoffel–perturbed mixed orthogonal polynomials, enabling both a transparent algebraic understanding and efficient computation. This factorization underpins total positivity and the spectral characterization of these discrete operators (Branquinho et al., 2024).

4. Step Recurrence in Topological and Ergodic Dynamics

Step recurrence sets and their generalizations are pivotal in ergodic theory and topological dynamics.

• $|\mathbf{n}|$8-Step Poincaré/Birkhoff Recurrence: A set $|\mathbf{n}|$9 is a $$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$0-step recurrence set if every measure-preserving (or minimal topological) system and every positive measure (or open) set $$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$1 has some $$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$2 with

$$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$3

(respectively, the intersection of $$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$4 under $$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$5 is nonempty). These generalize classical recurrence to higher steps (Huang et al., 2011).

• Combinatorial Characterizations: Various combinatorial families—$$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$6-step Poincaré sets, Birkhoff sets, $$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$7 sets, Nil$$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$8-Bohr$$\int x^k P_{\mathbf{n}}(x) \, d\mu_j(x) = 0, \quad k=0,1,\ldots,n_j-1, \; j=1,\ldots,r.$$9 sets (and their duals)—interrelate in strict inclusions, but coincide in characterizing regionally proximal relations and $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$0-step almost automorphic points for every neighborhood. This situates step recurrence within a hierarchy of recurrence types, foundational for the study of higher-order orbit structures.
• Regionally Proximal Relation of Order $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$1: In minimal systems, regionally proximal pairs of order $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$2 are fully captured by the presence of step recurrence in corresponding hitting-time sets for all neighborhoods, connecting combinatorial and dynamical formulations.

5. Higher-Order and Specialized Step Recurrences in Discrete Systems

• Explicit Multi-Step Recurrence Formulas: For standard two-step recurrences (e.g., Fibonacci type), closed-form product solutions exist, such as

$$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$3

with $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$4 rational in $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$5 and initial data. These yield exact identities for double factorial sequences and motivate generalizations to higher-step recurrences, with implications for total positivity and combinatorics (III, 2015).

• Applications to Molecular Motors and Statistical Physics: The detailed step-number dependence of recurrence and last-return probabilities in biased random walks illuminates behaviors in finite-length stochastic processes, such as the stepwise dynamics of molecular motors on microtubules, where critical step-length–bias relationships predict the disappearance of late-return peaks for processive biological transport (Mookerjee et al., 2024).

6. Synthesis: Interconnections and Distinctive Regimes

• Step recurrence is central in the analysis of discrete evolution in both deterministic (polynomial algebra, recurrence matrices) and stochastic (random walks, quantum walks) systems.
• The phenomenon is characterized by sharp critical thresholds (e.g., reinforcement parameter $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$6 in SRRW, growth rate exponent $$x p_m(x) = p_{m+1}(x) + \sum_{j=0}^r B_{m,j} p_{m-j}(x),$$7 in Rademacher walks, threshold bias in random walks, coin angle in DTQSWs) demarcating distinct recurrence regimes.
• Structural frameworks (e.g., multi-term recurrences in polynomials, bidiagonal factorization, combinatorial recurrence sets in dynamics) enable rigorous computation, classification, and, where possible, explicit solution of recurrence behavior.

Step recurrence thus embodies a unifying principle across combinatorics, stochastic processes, special function theory, and dynamical systems, providing both qualitative and quantitative tools for understanding return and memory phenomena in discrete mathematics and applied probability.