A constant-time algorithm for middle levels Gray codes

TL;DR

This paper presents an optimal-time algorithm for generating middle levels Gray codes, listing all relevant subsets with constant average time per generated set and linear space complexity.

Contribution

It introduces a novel algorithm that computes each new subset in constant time, improving upon previous algorithms with linear average time.

Findings

Each new set is generated in O(1) time on average.

The algorithm uses O(n) space.

It efficiently enumerates all middle levels Gray codes.

Abstract

For any integer $n\geq 1$ a middle levels Gray code is a cyclic listing of all $n$-element and $(n+1)$-element subsets of $\{1,2,\ldots,2n+1\}$ such that any two consecutive subsets differ in adding or removing a single element. The question whether such a Gray code exists for any $n\geq 1$ has been the subject of intensive research during the last 30 years, and has been answered affirmatively only recently [T. M\"utze. Proof of the middle levels conjecture. Proc. London Math. Soc., 112(4):677--713, 2016]. In a follow-up paper [T. M\"utze and J. Nummenpalo. An efficient algorithm for computing a middle levels Gray code. To appear in ACM Transactions on Algorithms, 2018] this existence proof was turned into an algorithm that computes each new set in the Gray code in time $\mathcal{O}(n)$ on average. In this work we present an algorithm for computing a middle levels Gray code in optimal time and space: each new set is generated in time $\mathcal{O}(1)$ on average, and the required space is $\mathcal{O}(n)$.