A Randomized Algorithm for Long Directed Cycle

TL;DR

This paper introduces a randomized algorithm for the Long Directed Cycle problem, leveraging existing algorithms for the k-Path problem to achieve efficient solutions with exponential time complexity.

Contribution

It establishes a method to solve the Long Directed Cycle problem using a black-box k-Path algorithm, improving the computational approach for directed cycles.

Findings

LDC can be solved in randomized time O*(4^k)

The approach reduces LDC to k-Path problem with a time bound

Provides a new connection between k-Path and LDC problems

Abstract

Given a directed graph $G$ and a parameter $k$, the {\sc Long Directed Cycle (LDC)} problem asks whether $G$ contains a simple cycle on at least $k$ vertices, while the {\sc $k$-Path} problems asks whether $G$ contains a simple path on exactly $k$ vertices. Given a deterministic (randomized) algorithm for {\sc $k$-Path} as a black box, which runs in time $t(G,k)$, we prove that {\sc LDC} can be solved in deterministic time $O^*(\max\{t(G,2k),4^{k+o(k)}\})$ (randomized time $O^*(\max\{t(G,2k),4^k\})$). In particular, we get that {\sc LDC} can be solved in randomized time $O^*(4^k)$.