Three time grids may appear (which do not depend on each other):
of the measurements.
where a piecewise linear approximation
of the control variable u(t)
is given by the values 
.
Also, any inequality constraint
of an optimal control problem
will only be satisfied
at the times of the control discretization grid.
.
Its purpose is to split up the whole integration
interval 
into a
sequence of smaller ones 
,
j=1,2,....
) also the control discretization grid
(in optimal control problems) always
must contain at least two grid points,
namely initial time 
and final time 
.