Optimizer burr phenomenon

[ghost]

In practical use, my expectation is a time series with a period of 0 followed by a slow increase, but the optimization result often becomes that during the expected time period of 0, there is not a uniform 0 result, but a trend of first increasing and then decreasing, with an increase of up to 0.1 values. I want to know what caused this?

[giaf]

Hi, it's difficult to say in general, you may want to check that:

• use another solver to check that indeed the numerical solution of your problem is what you expect.
• check that HPIPM did indeed converge and did so to an accurate enough solution, for example you could check the magnitude of the residuals.