


HIT_REDUCE_CONSTRAINTS Estimate regions from hyperplanes and a bounding polytope.
-------------------------------------------------------------------------
DESCRIPTION
-------------------------------------------------------------------------
[estimated_regions,adjacences,adjacences_hyp]=...
hit_reduce_constraints(sep_hyp, Bigregion,solver)
-------------------------------------------------------------------------
INPUT
-------------------------------------------------------------------------
sep_hyp{i,j} stores the hyperplane separating the i-th and j-th mode
datasets according to the formula sep_hyp{i,j}*[x 1]=0. By convention,
sep_hyp{i,i}=[0 .. 0] thus defining a fictious hyperplane.
Bigregion: bounding polytope containing the regions (usually the
regressor set).
solver: defines the solver to be used in mpt_solveLPs (called by
polyreduce_ext.m)
-------------------------------------------------------------------------
OUTPUT
-------------------------------------------------------------------------
estimated_regions(i): polytope defining the i-th region.
adjacences: each row is a pair (i,j), i<j indicating that the regions i
and j are adjacent The constraint i<j avoids storing both pairs (i,j) and
(j,i).
adjacences_hyp: each row contains the vector [w b] of the hyperplane
separating region i from region j (the regions are stored in the
corresponding row of adjacences) according to the formula [w b]*[x,1]'=0.