Hello, is there a way I can save the optimizer state and load it again while using the suggest-evaluate-register paradigm? Basically, my black box function is a large agent-based model that is executed from a separate script and takes around 3 hours to execute. Therefore, basically, this is the workflow I am looking for.
First,
Get a suggestion point from my optimizer and save it.
Run the model offline and get the target value.
Second,
Load previously saved optimizer
get the target value for the previous suggestion point
get the next suggestion and save the optimizer
Run the model offline with the suggested point and start from step 3 for the next runs.
dmehrab06
commented
1 year ago
Hello, is there a way I can save the optimizer state and load it again while using the suggest-evaluate-register paradigm? Basically, my black box function is a large agent-based model that is executed from a separate script and takes around 3 hours to execute. Therefore, basically, this is the workflow I am looking for.
First,
Second,