my workflow currently uses the Parameter Optimization Loop Start and end nodes.
I have to optimize a value based on two different parameters but the loop end node only allows for one.
Is it somehow possible to optimize towards two flow variables in the same Optimization Loop?
What is the outcome that you’re trying to optimize?
thanks for the question!
For one (what i am doing right now) I have a target value set up in the Loop Start Node
inside my loop I calculate a value containing a SalesVolume that should meet this target, i check this by taking the value I recieve from the calculation and subtracting it from the Target in a new column called Difference. This value should be minimized to meet the target as close as possible.
This part works as intended, what i need to do as well though is to maximize revenue. To maximize it, i have to multiply the SalesVolume by a Revenue Column.
So long story short, a possible optimization loop with two flow variables should get as close as possible to the given Target while also searching for the highest possible revenue.
Can you share the workflow with some example data?
It seems to me that incorporating the Pareto Ranking or Desirability Ranking nodes would be a better approach for multi-parameter optimization.
Hi @elsamuel sorry for the late response, it took a while to export the workflow. i could not simplify it unfortunately, the important stuff happens inside the Parameter Optimization Loop.
Loop Optimization.knar (221.4 KB)
The optimization loop needs to run in a specific way, thats why there is so much stuff happening beside it. The outer Group Loop Start loops by a period, the counting loop runs for the amount of rows contained in the period. That can basically all be ignored as it does not impact the actual optimization loop.
The important values in general are:
- EB_FZG is the revenue I want to multiply by (which should be maximized)
- The math formula contains the difference calculation between the goal value and what is originally calculated in the component before (FE Achievment w/o EV).
- The Parameter Optimization Loop End Node returns a column called Volume which is collected in the Loop end node after the Parameter Optimization Loop end.
Thats exactly the value I want to multiply the EB_FZG by and maximize it while keeping the difference between the calculated FE Achievment w/o EV as close as possible to the target.
Thank you for your help as always @elsamuel!! I hope you can do something with it
This topic was automatically closed 182 days after the last reply. New replies are no longer allowed.