I am new to Knime and would like to know if any of you have already done global optimizations with GenSa or similar algorithms?
Briefly to my problems. I am currently working on a concept for a dynamic pricing approach. It is now a matter of finding the maximum profit prices depending on the feature variables for the respective weeks of the year using the algorithm.
If these kinds of algorithms are not available in Knime, I would be very grateful if someone could explain how to integrate my model into R. The model used to estimate demand is XGBoost.
I would be very grateful for any advice, sample workflow, or suggestion.
In this thread I posted an example of a numeric prediction also using KNIME’s new XGBoost nodes. Maybe you could see if you could adapt that to your needs:
An example for XGBoost (Regression) can be found here:
A basic example of how to run a model in KNIME with R you might take a look at this:
About “GenSa algorithm” could you further elaborate on that what you are looking for.
Thank you very much for your detailed answer. I already appreciated the XGBoost model in Knime and optimized its parameters.
Now I need an algorithm that gives the maximum profit prices for each week in the past.
My target function is quite simple: Max margin = demand * price
(Since I have no costs in the Sample)
The starting price should simply be the last price. And the upper and lower bound is the minimum and maximum price available in the entire history.
Since I have in the estimated demand model other product prices and many more influence variables, I thought of the Generalized Simulated Annealing.
Hi JayJay -
Did you find a solution for this? I’m seeking to do the same type of thing and haven’t found much around finding local or global maxima in KNIME on top of a model.