Imaginary time propagation to find eigenfunctions

[semc]

Hi, I have been trying to use imaginary time propagation to get the ground state and excited states eigen function but the results I got is different from the analytical solution. I know that to get excited states, I should propagate 2 or more orthogonal functions depending on the number of excited states that you want. I just would like to check whether there is any requirement on the initial guess that you use to propagate. If I choose a random set of numbers to propagate, will it always contain my ground state?

 

[DrClaude]

If I choose a random set of numbers to propagate, will it always contain my ground state?

Most probably. After having found the ground state, any starting point to find the first excited state must be orthogonal to the ground state. If you start from a random guess, you need to first make it orthogonal to the ground state.

 

[chiro]

Hey semc.

What is your mathematical system you are solving for?