mpt_performance

MPT_PERFORMANCE Computes performance (i.e. sum of closed-loop costs) associated to a given controller

 [cost,XC] = mpt_performance(ctrl,gridpoints,Options)

 ---------------------------------------------------------------------------
 DESCRIPTION
 ---------------------------------------------------------------------------
 Grids the state-space into given number of points and for each state
 computes the closed-loop cost index. The returning argument is a sum of
 these costs.

 This function can be used to compare efficiency of low-complexity controllers
 with respect to optimal control strategy.

 USAGE:
   cost = mpt_performance(ctrl)
   cost = mpt_performance(ctrl,gridpoints)
   cost = mpt_performance(ctrl,gridpoints,Options)

 ---------------------------------------------------------------------------
 INPUT
 ---------------------------------------------------------------------------
 ctrl              - Explicit controller (MPTCTRL object)
 gridpoints        - number of grid points (if not provided, 30 is default)
 Options.verbose   - Level of verbosity
 Options.Pfinal    - polytope defining part of the state-space which should
                     be considered for cost computation. (only reasonable
                     if ctrlStruct.Pfinal is an empty polytope)

 Note: If Options is missing or some of the fields are not defined, the default
       values from mptOptions will be used

 ---------------------------------------------------------------------------
 OUTPUT
 ---------------------------------------------------------------------------
 cost   - contains "overall cost"
 XC     - a matrix which associates each state to a corresponding cost
          first n columns represent the state vector (x1,x2,...) and the last
          column is the cost obtained for this state
 X0     - set of feasible initial states (states which lie in
          ctrlStruct.Pfinal)
 Xinfeas - matrix of states for which the following holds:
             a) they feasible initial states (i.e. x \in Pfinal)
             b) no feasible closed-loop trajectory

 see also MPT_COMPUTETRAJECTORY, MPT_PLOTTIMETRAJECTORY