Improved Approximation for Longest Common Subsequence over Small Alphabets

TL;DR

This paper explores the approximability of the Longest Common Subsequence problem for small alphabets, showing the binary case is the most challenging and extending approximation results to larger constant-sized alphabets.

Contribution

It demonstrates that improving approximation ratios for binary LCS in subquadratic time implies similar improvements for all small alphabets, and extends existing results to all constant-sized alphabets.

Findings

Binary LCS is the hardest case for approximation.

Improved approximation algorithms are possible for all constant-sized alphabets.

Binary case remains the key challenge for better approximations.

Abstract

This paper investigates the approximability of the Longest Common Subsequence (LCS) problem. The fastest algorithm for solving the LCS problem exactly runs in essentially quadratic time in the length of the input, and it is known that under the Strong Exponential Time Hypothesis the quadratic running time cannot be beaten. There are no such limitations for the approximate computation of the LCS however, except in some limited scenarios. There is also a scarcity of approximation algorithms. When the two given strings are over an alphabet of size $k$, returning the subsequence formed by the most frequent symbol occurring in both strings achieves a $1/k$ approximation for the LCS. It is an open problem whether a better than $1/k$ approximation can be achieved in truly subquadratic time ($O(n^{2-\delta})$ time for constant $\delta>0$). A recent result [Rubinstein and Song SODA'2020] showed that a $1/2+\epsilon$ approximation for the LCS over a binary alphabet is possible in truly subquadratic time, provided the input strings have the same length. In this paper we show that if a $1/2+\epsilon$ approximation (for $\epsilon>0$) is achievable for binary LCS in truly subquadratic time when the input strings can be unequal, then for every constant $k$, there is a truly subquadratic time algorithm that achieves a $1/k+\delta$ approximation for $k$-ary alphabet LCS for some $\delta>0$. Thus the binary case is the hardest. We also show that for every constant $k$, if one is given two strings of \emph{equal} length over a $k$-ary alphabet, one can obtain a $1/k+\epsilon$ approximation for some constant $\epsilon>0$ in truly subquadratic time, thus extending the Rubinstein and Song result to all alphabets of constant size.