Talk:MPC


 * Is the PI term in the linear example really a PI or more of a PD model?
 * What about integrating further into the future? Currently the models focus on one time step in the future
 * How do you choose the decay constant,T, for the desired output? Why not choose it as zero to get the fastest response?

The topic is very important but difficult to illustrate with a simple simulator. Ideally this simulator would include:
 * mutliple time points in the future
 * explicit statement of constraints
 * maybe a variable input (variable in a known way--say a sin or cos function)