[Paper Review] New Lower Bounds on Sizes of Permutation Arrays

Lizhen Yang, Kefei Chen+1 more|ArXiv.org|Jan 25, 2008

graph theory and CDMA systems3 citations

TL;DR

This paper presents new lower bounds on the maximum size $ P(n,d) $ of permutation arrays (PAs) of length $ n $ and minimum Hamming distance $ d $, using graph-theoretic frameworks and ball-covering intersection techniques. It improves upon the classical Gilbert-Varshamov bound and establishes tighter bounds for specific $ n $ and $ d $, including $ P(11,5) \geq 95040 $ and $ P(11,6) \geq 15840 $, leveraging recursive constructions and combinatorial bounds on derangements and permutation supports.

ABSTRACT

A permutation array(or code) of length $n$ and distance $d$, denoted by $(n,d)$ PA, is a set of permutations $C$ from some fixed set of $n$ elements such that the Hamming distance between distinct members $\mathbf{x},\mathbf{y}\in C$ is at least $d$. Let $P(n,d)$ denote the maximum size of an $(n,d)$ PA. This correspondence focuses on the lower bound on $P(n,d)$. First we give three improvements over the Gilbert-Varshamov lower bounds on $P(n,d)$ by applying the graph theorem framework presented by Jiang and Vardy. Next we show another two new improved bounds by considering the covered balls intersections. Finally some new lower bounds for certain values of $n$ and $d$ are given.

Motivation & Objective

• To improve existing lower bounds on $ P(n,d) $, the maximum size of permutation arrays with length $ n $ and minimum Hamming distance $ d $.
• To develop tighter bounds than the classical Gilbert-Varshamov bound using graph-theoretic and combinatorial techniques.
• To establish new explicit lower bounds for specific values of $ n $ and $ d $, especially for $ n=11 $.
• To explore structural properties of permutation arrays via covered balls and support intersections.

Proposed method

• Applying the graph theorem framework of Jiang and Vardy to refine the Gilbert-Varshamov lower bounds on $ P(n,d) $.
• Analyzing intersections of covered balls (Hamming balls of radius $ d-1 $) to derive improved bounds.
• Using recursive constructions based on fixing positions in permutations and projecting to smaller symmetric groups.
• Defining and analyzing $ ψ^s_i $ and $ ψ^t_j $ mappings to preserve minimum distance after dimension reduction.
• Leveraging known results on mutually orthogonal Latin squares and derangement counts to bound $ P(n,d) $.
• Deriving recursive inequalities such as $ P(n-1,d-3) \geq P(n,d) $ and $ P(n-1,d-2) \geq \frac{2}{n}P(n,d) $.

Experimental results

Research questions

• RQ1 Can the Gilbert-Varshamov lower bound for $ P(n,d) $ be improved using graph-theoretic methods?
• RQ2 How do intersections of covered balls in permutation spaces affect lower bounds on $ P(n,d) $?
• RQ3 What tighter explicit lower bounds can be derived for $ P(n,d) $ when $ n $ and $ d $ are specific, such as $ n=11 $?
• RQ4 What recursive relationships exist between $ P(n,d) $ and $ P(n-1,d-k) $ for small $ k $?
• RQ5 How do combinatorial structures like derangements and orthogonal Latin squares influence bounds on $ P(n,d) $?

Key findings

• The paper establishes $ P(11,5) \geq 95040 $, improving upon the prior bound of 60940.
• It proves $ P(11,6) \geq 15840 $, exceeding the previous bound of 9790.
• For prime powers $ q $, $ P(q,q-4) \geq (q+1)q(q-1) $, which is tighter than previous bounds in certain cases.
• The bound $ P(q,q-3) \geq 2q(q-1) $ is derived and shown to be superior to earlier results under specific modular conditions on $ q $.
• The recursive inequality $ P(n-1,d-3) \geq P(n,d) $ is proven for $ n \geq d > 3 $, enabling inductive lower bound generation.
• The bound $ P(q-1,q-6) \geq 2(q+1)(q-1) $ is established using the same recursive framework.