MPT_BLOCKINGMATRICES Constructs matrices for the CFTOC problem for move blocking strategies
[Matrices]=mpt_blockingMatrices(Matrices,sysStruct,probStruct,Options)
---------------------------------------------------------------------------
DESCRIPTION
---------------------------------------------------------------------------
Constructs cost and constraint matrices for the finite time constrained
optimal control problem on linear time-invariant systems with quadratic
cost (2-norm) in the case where inputs or their differences are fixed to be
constant over a certain number of steps during the prediction horizon N.
Matrices received from function mpt_constructMatrices are compressed
depending on probStruct.inputblocking and probStruct.deltablocking.
Degrees of freedom are reduced from full degrees of freeedom m * N (m=number
of inputs, N=prediction horizon) in the non blocking case to m * M (M < N,
M independent decision variables in blocked input sequence) in the case where
inputs are their differences are fixed to be constant during the prediction
horizon.
---------------------------------------------------------------------------
INPUT
---------------------------------------------------------------------------
Matrices NORM=2
--------
G,W,E,H,F,Y,Cf,Cx,Cc - matrices of the problem, i.e.
sysStruct - System structure in the sysStruct format
probStruct - Problem structure in the probStruct format
- NORM=2
- horizon
Prediction horizon N; How many time steps are considered.
- inputblocking
Fix inputs u_k to be constant over a certain number of steps
in the prediction horizon.
inputblocking=[n1 n2 ... nk], \sum_{n1}^{nk} = N.
Entries define how many consecutive inputs are fixed to be
constant. Sum of all entries has to be equal the prediction
horizon N.
Example: N = 5, inputblocking = [1 4]
First predicted input is independent and the
next 4 predicted inputs are fix to be constant,
i.e. u1, u2=u3=u4=u5.
- deltablocking
Fix difference u_k - u_{k+1} to be constant over a certain number
of steps in the prediction horizon.
deltablocking = [1 ... nk ... N], 1 < .. < .. nk < .. < N
Entries defines wich inputs are independent (=decision variables),
inputs in between are interpolated (=constant differences).
Example: N = 5, deltablocking = [1 5]
First input u1 and last input u5 are independent,
inputs in between are interpolated, i.e.
u1-u2 = u2-u3 = u3-u4 = u4-u5.
Note: If Options is missing or some of the fields are not defined, the default
values from mptOptions will be used
---------------------------------------------------------------------------
OUTPUT
---------------------------------------------------------------------------
Matrices NORM=2
--------
Compressed matrices for fixed inputs or fixed differences in
prediction horizon.
G,W,E,H,F,Y,Cf,Cx,Cc - matrices of the problem, i.e.
---------------------------------------------------------------------------
LITERATURE
---------------------------------------------------------------------------
"Move Blocking Strategies in Receding Horizon Control"
R. Cagienard, P. Grieder, E. C. Kerrigan and M. Morari, 2004, submitted
check http://control.ee.ethz.ch for latest info
see also MPT_CONSTRUCTMATRICES
Copyright is with the following author(s):
(C) 2004 Raphael Cagienard, Automatic Control Laboratory, ETH Zurich,
(C) 2003 Pascal Grieder, Automatic Control Laboratory, ETH Zurich,
cagienard@control.ee.ethz.ch