1: \begin{abstract}

2: %Variable steps implicit-explicit multistep methods  for  PDEs have been presented in \cite{WR:08}, where  the zero-stability  is studied for ODEs;

3: %however, the stability analysis  still remains an open question for PDEs.

4: 

5: In this work, we analyze the three-step backward differentiation formula (BDF3) method

6:  for solving the Allen-Cahn equation on variable grids.

7: For BDF2 method, the discrete orthogonal convolution (DOC) kernels are positive, the stability and convergence analysis are well established in [Liao and  Zhang, \newblock  Math. Comp., \textbf{90} (2021) 1207--1226; Chen,  Yu, and  Zhang, \newblock SIAM J. Numer. Anal., Major  Revised]. However, the numerical analysis for BDF3 method with variable steps seems to be highly nontrivial, since the DOC kernels are not always positive.

8: By developing a novel spectral norm inequality,

9: the unconditional stability  and convergence are rigorously proved under the updated step ratio restriction $r_k:=\tau_k/\tau_{k-1}\leq 1.405$

10: (compared with $r_k\leq 1.199$ in [Calvo and Grigorieff, \newblock  BIT. \textbf{42} (2002) 689--701]) for BDF3 method.

11: Finally, numerical experiments are performed  to illustrate the theoretical results.

12: To the best of our knowledge, this is the first theoretical analysis of variable steps BDF3 method for the Allen-Cahn equation.

13: \end{abstract}