We want to couple an in-house flow solver with an in-house optimization code, both written in Matlab as m-files (both have many subroutines and a main function). This is to optimize some geometric parameters.
Once a single flow solution case is crashed (e.g. in one of solver subroutines), the whole optimization process is crashed (the main flow solver m-file is defined as a function in the main optimization m-file).
Our aim is to continue the flow solution cases after one case is crashed (e.g. due to NaN error, etc.).
Is there a way to accomplish this without adding complex and numerous controls to the solver code? (e.g. to terminate the main solver m-file without terminating the main optimization m-file).
By the way crashes occur in the subroutines of the flow solver, not in the main flow solver m-file.