[Paper Review] Partition Parameters for Girth Maximum (m, r) BTUs

Vivek S. Nittoor, Reiji Suda|arXiv (Cornell University)|Dec 31, 2012

Error Correcting Code Techniques6 references3 citations

TL;DR

This paper establishes a mathematical framework for identifying optimal partition parameters β₁, β₂, ..., βᵣ₋₁ in P₂(m) such that the girth-maximum (m,r) Balanced Tanner Unit (BTU) lies within the family Φ(β₁, β₂, ..., βᵣ₋₁). It derives a systematic search strategy based on symmetric permutation trees and cycle index theory, proving that optimal partitions can be constructed via recursive scaling of base partitions, enabling efficient search for BTUs with maximal girth.

ABSTRACT

This paper describes the calculation of the optimal partition parameters such that the girth maximum (m, r) Balanced Tanner Unit lies in family of BTUs specified by them using a series of proved results and thus creates a framework for specifying a search problem for finding the girth maximum (m, r) BTU. Several open questions for girth maximum (m, r) BTU have been raised.

Motivation & Objective

• To determine the optimal partition parameters β₁, β₂, ..., βᵣ₋₁ ∈ P₂(m) such that the girth-maximum (m,r) BTU belongs to the family Φ(β₁, β₂, ..., βᵣ₋₁).
• To establish a mathematical foundation for constructing (m,r) BTUs with maximum girth using compatible permutations on symmetric permutation trees.
• To create a structured search problem for identifying girth-maximum (m,r) BTUs by leveraging recursive partition scaling.
• To raise open questions on the theoretical limits and computational complexity of finding such optimal BTUs.

Proposed method

• Uses symmetric permutation trees S_PT{m} to model the structure of (m,r) BTUs, ensuring all permutations are compatible and non-repeating across levels.
• Defines partition components βᵢ ∈ P₂(m) as the cycle structure between consecutive permutations pᵢ and pᵢ₊₁ in the tree traversal.
• Applies cycle index theory and permutation group compatibility to ensure that pᵢ₊₁ ∉ C(p₁, ..., pᵢ), maintaining non-isomorphic BTU construction.
• Proposes a recursive algorithm to generate optimal partitions: for given k and r, scale base partitions βᵢ = {kⁱ} with multiplicity k^{r-1-i} to form βᵢ ∈ P₂(m) where m = b·k^{r-1}.
• Constructs girth-maximum BTUs by hierarchical search: first find optimal (b·k,2) BTU, then (b·k²,3), up to (b·k^{r-1},r), using the same partition pattern.
• Employs the condition that b is minimized in the factorization m = k^{r-1}·b to ensure minimal redundancy and maximal girth potential.

Experimental results

Research questions

• RQ1 What is the maximum attainable girth for a (m,r) BTU, and what structural conditions enable it?
• RQ2 How can the search space for girth-maximum (m,r) BTUs be optimally parameterized using partition families Φ(β₁, ..., βᵣ₋₁)?
• RQ3 What is the computational complexity of identifying a girth-maximum (m,r) BTU under the proposed partition framework?
• RQ4 Can a girth-maximum (m,r) BTU be constructed recursively by extending smaller optimal BTUs with consistent partition patterns?
• RQ5 What role do symmetric permutation trees and compatible permutation sequences play in ensuring maximal girth?

Key findings

• Any (m,r) BTU is isomorphic to an element of Φ(β₁, β₂, ..., βᵣ₋₁) for some βᵢ ∈ P₂(m), establishing a complete parameterization of all non-isomorphic BTUs.
• Optimal partitions βᵢ ∈ P₂(m) for girth-maximum (m,r) BTUs are constructed as βᵢ = ∑_{j=1}^{k^{r-1-i}} {b·kⁱ} = b·k^{r-1} = m, where m = b·k^{r-1} and b is minimized.
• The framework enables a hierarchical search: girth-maximum (m,r) BTUs can be built by successively finding optimal (b·k²,3), (b·k³,4), ..., (b·k^{r-1},r) BTUs.
• For r=3, the search reduces to finding p₃ ∈ S_{b·k²} such that the labeled BTU {p₁, p₂, p₃} achieves maximum girth, given p₁ = I_{b·k²} and p₂ ∈ Ψ(β₁) with β₁ = ∑_{j=1}^{k} {b·k} = b·k².
• An algorithm is provided to generate optimal partitions: initialize βᵢ = {kⁱ}, scale by k for each prior index, and update m = k·m iteratively for i = 1 to r−1.
• The conjecture holds that constructing a girth-maximum (k²,3) BTU suffices as a template for building higher-order girth-maximum (m,r) BTUs via recursive partition scaling.