I have a control system problem which requires to solve quadratic programming to minimize quadratic cost function under inequality constraints.
The aim is to find control input increments which minimize the cost function.
For the first case, my inputs are relatively small values like 20 [N] and 150 [Nm] and the error between the current state and reference state is calculated by the terms in parenthesis where the reference values have 0.1 [m] increment.
When I map the inputs to new sets of inputs having relatively large values like 9810 [rpm^2] for nominal value and 25000 [rpm^2] for maximum value, quadratic programming outputs really small values which does not affect the system as much as it requires. So, it stays where it started in equilibrium point.
Is there anything that I should consider about the magnitudes of the quantities used in cost function or quadratic programming?