Adding a constraint for velocity increasing with depth

[ariellellouch]

Geologically (especially at large scales), we often want the velocity models to increase with depth. The current model parametrization doesn't really allow to do that (unless layers are mutually exclusive). Could it be possible to add a flag for this constraint, so that for the layer number n+1, the velocity will be between [velocity of layer n, max allowed velocity of layer n+1]?

Thanks again!

[keurfonluu]

Hi @ariellellouch,

That was actually planned, I just didn't have time to implement such feature. But basically, right now, it's possible to do it by:

• manually initializing the population with only increasing velocity models (yet still random)
• adding a penalty term for models that do not satisfy that constraint (via argument extra_terms in configure)

But it will definitely be better to have a boolean option to automate that.

[ariellellouch]

Thanks! I am a bit scared to play with the extra_terms without an example, but will give it a go! :)

Edit: Quick and dirty solution which assures Vs is going up with depth (I'm sure this can be made prettier and not depend on the structure of x, but I am not a good enough coder). Adding extra_terms=inc_misfit works for me.

def inc_misfit(x): nel = int((np.size(x)-2)/3) return 999999.9 np.any(np.diff(x[nel:2nel+1])<0)

[keurfonluu]

As long as it works!

I added an option increasing_velocity in configure that initializes the population with increasing velocity models, and adds a penalty term to reject generated models given the constraint:

model.configure(
    optimizer="cpso",  # Evolutionary algorithm
    misfit="rmse",  # Misfit function type
    optimizer_args={
        "popsize": 10,  # Population size
        "maxiter": 100,  # Number of iterations
        "workers": -1,  # Number of cores
        "seed": 0,
    },
    increasing_velocity=True,
)

You can update to evodcinv v2.1.1: pip install evodcinv -U.

[ariellellouch]

Fantastic, thanks again and sorry for bugging you :)