I need help with this optimization problem which is either not getting solved at all or is taking a copious amount of time.

I am trying to find optimized input to an RNN (GRU) model of a process plant, which will minimize the difference between the output and desired output. I am using obj = (model_output(x) - desired_output(x))**2

If I use 1) sol = minimize(objective,x0, method = 'SLSQP', bounds = bnds, options = {'disp' : True}) It just gives back x0

2) sol = differential_evolution(objective,bnds) is taking extremely large amount of time (Several minutes for 1 x. If I use it iteratively, it takes several hours)

P.S. I mentioned recursively since x is an input over 20 timesteps, then I need to move from 1:20 indices to 2:21 and so on till 381:400 timesteps for my problem. So I am running this operation recursively using for loop.

Also bounds are simple bounds to input like x should be between (60,80).

  • $\begingroup$ I tried using hyperopt for Bayesian Optimization but I am not able to figure out how to give the "space" variable for a series of input variables to be optimized (in my case, a series of 20 input variables due to 20 timesteps). $\endgroup$ – Sam123 Jan 19 at 19:36

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.