Python Forum
Optimise experiment control parameters
Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Optimise experiment control parameters
#1
I say chaps,
I have an experiment controlled by 6 parameters, hooked up to a Raspberry Pi. I want to run a loop: set parameters, do experiment, get result, estimate better parameters based on previous history, do experiment.......
How can I do this in Python? I've trawled the Web but the optimise routines seem to presume some equation to find the maximum of, or some purely mathematical situation, not a physical contrivance. I obviously haven't found the optimum Google search terms!
I could optimise the parameters two at a time, first coarse tune , then fine tune then finer tune.
Suggestions pointing to some bit of code already out there would be most welcome.
TTFN
rickticktock
Reply
#2
You can do random search, run for loop and choose parameters in multidimensional space randomly, do experiment, and choose the best case; another approach is to use simulated annealing method, it is almost the same, but takes into account local estimations of probability where the optimal solution is likely existing. Everything depends on how much time do you need to do one experiment? Searching in 6D might be hard...
Reply
#3
Thanks scidam. A 2d search could be the way to go, then another 2d search, then another 2d search. Then start again. The physical problem changes slowly anyway, it's literally a question of keeping on top of it.
TTFN
Reply


Possibly Related Threads…
Thread Author Replies Views Last Post
  Optimise multiply for loop in Python KenBCN 4 428 Feb-06-2024, 06:48 PM
Last Post: Gribouillis
  Compare response and name list in experiment knoxvillerailgrind 3 2,174 Jul-26-2020, 12:23 PM
Last Post: deanhystad
  Bode plot from time series experiment data discus 4 7,209 Jun-20-2020, 07:46 AM
Last Post: discus
  lab experiment / encryption pythan 0 2,436 Jun-09-2018, 07:19 PM
Last Post: pythan

Forum Jump:

User Panel Messages

Announcements
Announcement #1 8/1/2020
Announcement #2 8/2/2020
Announcement #3 8/6/2020