Weighted Motzkin Paths in Combinatorics

• Weighted Motzkin Paths are generalized lattice paths defined by assigning weights to each step (up, level, down) to encode combinatorial structures and applications.
• They incorporate generating functions, continued fractions, and Riordan arrays to capture detailed enumerative statistics and establish bijections with other path families.
• The framework applies to enumerative combinatorics, statistical mechanics, and orthogonal polynomial theory, providing insights into asymptotic regimes and algorithmic analysis.

A weighted Motzkin path is a generalization of the classical Motzkin path, where each step (up, level, down, and in some models, vertical or horizontal) is assigned a weight drawn from a specified parameter sequence or function. The total weight of a path is the product of its step weights. Weighted Motzkin paths serve as combinatorial representations for diverse phenomena in enumerative combinatorics, probability, statistical mechanics, and bijective correspondence with algebraic objects such as orthogonal polynomials. Their study incorporates continued fractions, Riordan arrays, pattern avoidance, and intricate bijections to other weighted path families.

1. General Definition and Step Weights

A Motzkin path of length $n$ is typically a lattice path from $(0,0)$ to $(n,0)$ in $\mathbb{Z}^2$ staying nonnegative in the $y$-coordinate, composed of three types of steps: up $(1,1)$, level $(1,0)$, and down $(1,-1)$—i.e., $(U,H,D)$. More generally, the weighted Motzkin path formalism associates a weight to each step, which may depend on the type, position, or height:

• For classical weighted Motzkin paths (Sun et al., 2022), assign weight $1$ to each up-step, weight $a$ to each horizontal (level) step, weight $b$ to each vertical step (e.g., $(0,-1)$), and weight $c$ to each down-step.

The generating function encoding all weighted Motzkin paths of length $n$ is defined as

$$G_n(a,b,c) = \sum_{P:\,|P|=n} w(P),\qquad G(a,b,c;\,x) = \sum_{n\ge0}G_n(a,b,c)\,x^n$$

with $w(P)$ the total path weight. The functional equation derived by first-step decomposition is

$$G(a,b,c;\,x) = \frac{1 - a\,x \;-\;\sqrt{(1 - a\,x)^2 \;-\;4\,x\,(b+c\,x)}}{2\,x\,(b + c\,x)}$$

Weight assignments may be specialized, for instance, as height-dependent or governed by recursive definitions (e.g., Fibonacci, $q$-analogue, or linear in height).

2. Pattern-Avoidance Models and Bijections

The introduction of additional step types (e.g., vertical steps), together with pattern-avoidance constraints, enriches the weighted Motzkin path paradigm:

• An $(a,b,c)$-G-Motzkin path allows steps $u=(1,1)$ (up), $h=(1,0)$ (horizontal), $v=(0,-1)$ (vertical), and $d=(1,-1)$ (down), with respective weights (Sun et al., 2022).
• A path is $\mathbf{uvu}$-avoiding if the subsequence $uvu$ does not appear. The generating function for $\mathbf{uvu}$-avoiding $(a,b,c)$-G-Motzkin paths, $\mathcal G_{n}^{\rm uvu}(a,b,c)$, is:

$$G^{\rm uvu}(a,b,c;\,x) = \frac{(1 - a x)(1 + b x)-\sqrt{(1 - a x)^2(1 + b x)^2-4x(1 + b x)(b + c x)}}{2x\,(b + c x)}$$

In the specialization $c=b^2$, this coincides with the generating function for $(a,b)$-Schröder paths, facilitating explicit bijections:

$$G^{\rm uvu}(a,b,b^2;\,x) = \frac{1 - a x - \sqrt{(1 - a x)^2-4b x}}{2b x}$$

There are weight-preserving bijections between restricted $(a,b,c)$-G-Motzkin paths and weighted Schröder or Dyck path families, central for combinatorial enumerations and explaining algebraic identities among their counting sequences (Sun et al., 2022).

3. Connection to Riordan Arrays and Enumerative Statistics

Weighted Motzkin paths encode rich level-wise statistic arrays. For $\mathbf{uvu}$-avoiding G-Motzkin paths, various enumerative statistics are computed by exploiting Riordan-array structures:

• $U_{n,i}$, $V_{n,i}$, $D_{n,i}$, $H_{n,i}$, $P_{n,i}$ count, respectively, up, vertical, down, horizontal steps, or points at height $i$,
• These statistics are cached in Riordan arrays of the form $(d(x), h(x))$ with analytic generating-function representations.

For example, the up-step array is

$$U_{n,i} = [x^{\,n+1}]\, x^{\,i+1}\,\frac{S(x)^{2i+3}}{1+x}$$

where $S(x)$ is the large-Schröder generating function.

These enumerative results provide closed-form double sums and explicit matrix representations for combinatorial statistics, further linking to classical objects such as the Catalan triangle or Narayana numbers in specific parameter selections (Sun et al., 2022).

4. Algorithmic Analysis and Permutations via Area

Weighted Motzkin path formulations are central in resolving enumerative problems such as counting permutations of fixed total displacement (Bärtschi et al., 2016):

• The area statistic for a Motzkin path (the sum of step heights) precisely matches half the total displacement of a permutation under a classic bijection.
• Step weights tied to height: up/down steps at height $h$ get weight $h$, horizontal steps at $h$ get weight $2h+1$.

The total number of permutations of displacement $2d$ for $n$ elements is

$D(n,d) = \sum_{mz\in MZ(n,d)} \omega(mz)$

with $\omega(mz)$ as the product of step weights.

Efficient dynamic programming and MCMC algorithms have been built for enumeration and uniform sampling of weighted Motzkin paths with fixed area, using structures such as building sequences $a = (f_0, p_1, f_1, ..., p_h, f_h)$, where $f_i$ is the number of horizontal steps at height $i$, and $p_i$ the number of peaks at $i$ (Bärtschi et al., 2016).

5. Weighted Motzkin Paths in Orthogonal Polynomial and Statistical Mechanisms

Weighted Motzkin paths encode the moments and combinatorial models for families of orthogonal polynomials, notably the $q$-Hermite, $q$-Charlier, and $q$-Laguerre polynomials (Josuat-Vergès et al., 2010). The step weight sequences correspond to the three-term recurrence relations of these polynomials:

• $p_h$, $q_h$, $r_h$ are mapped to coefficients of the recurrence,
• Motzkin path enumeration with these weights coincides with combinatorial objects such as matchings (Touchard-Riordan), set partitions, and permutation crossings.

The generating functions often admit a J-fraction (continued fraction) representation:

$$M(z) = \cfrac{1}{1 - q_0\,z - \cfrac{p_0\,r_1\,z^2}{1 - q_1\,z - \cfrac{p_1\,r_2\,z^2}{1 - q_2\,z - \cdots}}}$$

These analytic forms tie directly to hypergeometric expressions for moments and counting formulas pertinent to partition statistics (Josuat-Vergès et al., 2010).

Weighted Motzkin paths further model physical processes such as the totally asymmetric exclusion process (TASEP) with colored step weights encoding domain wall configurations and serve as transfer-matrix enumerations in statistical mechanics (Woelki, 2013).

6. Specializations, Generalizations, and Applications

Weighted Motzkin path generating functions interpolate smoothly between established combinatorial sequences by specialization:

• $(a,b,c) = (1,1,1)$ yields Motzkin numbers.
• $(a,b,c) = (1,1,0)$ reduces to Catalan numbers.
• $(a,b,c) = (1,1,1^2)$ gives Schröder numbers (Sun et al., 2022).

Pattern avoidance ($uvu$, $uvv$, $uu$) produces bijective correspondences among path families and equinumerosity results. The functional equations persist in generalizations such as periodic Motzkin paths, weighted compositions, and infinite automata, with explicit continued fraction representations for enumerative or moment generating functions (Gan et al., 2022, Castro et al., 2013).

Weighted Motzkin paths underpin the combinatorial identities relating Motzkin, Catalan, Schröder, and Fine numbers, and connect via Riordan arrays and determinant identities to algebraic formulas for various triangles (Catalan, Shapiro, Riordan) (Sun et al., 2013, Chen et al., 2016).

7. Asymptotic Regimes and Analytic Geometry

Recent advances address weighted Motzkin paths with step weights varying linearly with height ("balanced regimes"), connecting combinatorics to the Pearson structure of PDEs, saddlepoint approximations, and large-deviation principles (Omelchenko, 24 Jan 2026):

• The exponential generating function satisfies a Pearson-type PDE, solvable by the method of characteristics.
• Moving algebraic singularities control local and global behavior, with explicit limit theorems: local Gaussian regime, large-deviation rate function $I(u)$, and uniform Daniels-type saddlepoint formulas for the probability of endpoint configurations.

These analytic tools generalize to tridiagonal recurrences and block-structured birth-death processes.

---

Weighted Motzkin paths form a central structure in modern enumerative combinatorics and its applications to probability, algebra, and integrable systems, admitting explicit combinatorial, algebraic, and analytic characterization. Their flexible framework enables bijective enumeration, powerful generating function technology, cross-links to orthogonal polynomials, and rigorous asymptotic analysis. The literature cited above provides exhaustive formulas, bijections, and computational tools for their study and specialization.