# Hamiltonicity below Dirac's condition

## TL;DR

This paper develops fixed-parameter algorithms for the Hamiltonian cycle problem in graphs where the minimum degree condition is slightly relaxed from Dirac's classical theorem, extending the problem's tractability.

## Contribution

It introduces the first fixed-parameter algorithms for Hamiltonicity below Dirac's bound, with optimal exponential-time complexity and polynomial kernelization.

## Key findings

- Algorithms run in $c^k imes n^{O(1)}$ time, optimal under ETH.
- A polynomial kernel with $O(k)$ vertices is achievable.
- Extends tractability of Hamiltonian cycle problem to near-Dirac graphs.

## Abstract

Dirac's theorem (1952) is a classical result of graph theory, stating that an $n$-vertex graph ($n \geq 3$) is Hamiltonian if every vertex has degree at least $n/2$. Both the value $n/2$ and the requirement for every vertex to have high degree are necessary for the theorem to hold. In this work we give efficient algorithms for determining Hamiltonicity when either of the two conditions are relaxed. More precisely, we show that the Hamiltonian cycle problem can be solved in time $c^k \cdot n^{O(1)}$, for some fixed constant $c$, if at least $n-k$ vertices have degree at least $n/2$, or if all vertices have degree at least $n/2-k$. The running time is, in both cases, asymptotically optimal, under the exponential-time hypothesis (ETH). The results extend the range of tractability of the Hamiltonian cycle problem, showing that it is fixed-parameter tractable when parameterized below a natural bound. In addition, for the first parameterization we show that a kernel with $O(k)$ vertices can be found in polynomial time.