Batch mode exhaustive

[Benjamin Mewes]

Dear all,

right now I try to calibrate my model using observed data. In order to find a fitting set of parameters I want to vary them by using a batch experiment. I use the exhaustive method right now to find my parameters. What is strange:

When I calibrate them parameter by parameter (for each calibration a few runs based on the parameter space) it calculates different goodness-of-fit values. If I try to calibrate them with more than one parameter to explore, nothing (!) changes.

Working:

experiment Infiltration_mitTH_batch type:batch repeat: 2 keep_seed:true until: (cycle = 100){
        parameter 'ks' var:ks min: 300.0 max: 700.0 step: 10.0;
       method exhaustive maximize: KGE ;
}


Not working:
experiment Infiltration_mitTH_batch type:batch repeat: 2 keep_seed:true until: (cycle = 100){

        parameter 'Qs' var: gQs min: 0.1 max: 0.8 step: 0.1;
        parameter 'geta' var:geta min: 0.1 max: 0.8 step: 0.1;
        parameter 'ks' var:ks min: 300.0 max: 700.0 step: 10.0;
       
        method exhaustive maximize: KGE ;
}

To be honest, I don't understand why it works with one parameter, but doesn't work with more than one. If I use the second batch definition, it states that the values are changed, but the results remain always the same.

Moreover, I'd like to get the best fit of my model to a pre-defined value "KGE = 1", is there an implemented method which doesn't just minimizes or maximizes a value but optimises towards a defined value?

Thanks in advance!

[Benjamin Mewes]

Got it working, very strange as a clean rebuild of GAMA did help eventually.