# Eigenvalue problem with several domains

**URL:** https://community.freefem.org/t/eigenvalue-problem-with-several-domains/60
**Category:** General Discussion
**Created:** [May 28, 2019, 10:29am UTC](https://community.freefem.org/t/eigenvalue-problem-with-several-domains/60 "2019-05-28T10:29:55Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![GaelleB](https://avatars.discourse-cdn.com/v4/letter/g/34f0e0/32.png) [@GaelleB](https://community.freefem.org/u/GaelleB)
#### Post date: [May 28, 2019, 10:29am UTC](https://community.freefem.org/t/eigenvalue-problem-with-several-domains/60/1 "2019-05-28T10:29:55Z")

</div>

Hello,

I solved an eigenvalue problem on a domain D= D1 U D2 using the matrix construction as explained in the Freefem documentation. Each domain is triangulated and the meshes are adapted with the corresponding metric. My operators are symmetric and I know that theoretically the zero eigenspace is three-dimensional.

My operator does not have the same expression in each domain. And even if my Freefem code was computed, numerically I do not have a triple zero eigenvalue. I think this is because the basis is not the same for D1 and D2, and it is not taken into account in the assembling of matrices. I did not have any warning in my code because I was working with several domains.

We found one method to solve this problem with C++, and we think that maybe it can be added on Freefem software.

I can send more details in a PDF file if you are interested.