1: \begin{abstract} The gentlest ascent dynamics (GAD) 

2:  ({\it Nonlinearity}, vol. 24, no. 6, p1831, 2011)  is a continuous time  dynamics

3:  coupling both the position and the direction variables to efficiently locate the

4:  saddle point with a given index. 

5:  These  saddle points play important roles

6:  in the   activated process of the randomly perturbed 

7:  dynamical systems.

8:  For index-1 saddle points in non-gradient systems,  the GAD

9:  requires two direction variables to approximate the  eigenvectors of the Jacobian matrix and its transpose, respectively,

10:  while in the gradient systems, these two directions collapse to be the single min mode of the Hessian matrix.

11:  In this note,  we present a simplified GAD which only  needs one direction variable  even for non-gradient systems.

12:  This new method not only  reduces computational cost for directions by half,

13:  but also  can avoid inconvenient operations on  the transpose of Jacobian matrix.

14:   We prove  the same convergence property for the simplified GAD

15:    as for the original GAD.  The motivation of our simplified GAD is

16:  its formal analogy  to the Hamiltonian dynamics  governing the exit dynamics

17: when the system is perturbed by small  noise.

18: Several non-gradient examples are presented to demonstrate our method,

19: including the two dimensional models and

20: the Allen-Cahn equation in the presence of shear flow.

21: 

22: 

23: \end{abstract}

24: