🚀 go-pugleaf

In article <Dec.21.13.36.30.1992.21577@gandalf.rutgers.edu> rgonzal@gandalf.rutgers.edu (Ralph Gonzalez) writes:
>jwiegand@moe.eng.temple.edu (James Wiegand) writes:
>
>I'm not an expert in this area, but I think the only general-purpose
>approach for nonlinear control systems is the brute-force approach.

Like any other optimization problem, there are much more intelligent
means for solving optimal control problems than brute force.  Take
a look at "Computational Methods in Optimization," Polak, 1971.


>In the 60's, King-Sun Fu and others took the approach of partitioning
>state space, and "learning" the (sub) optimal control action for
>each element of the partition, on-line. The resulting "lookup-table"
>allows the controller to select the best control action for each
>point in state space. 

There was another approach similar to this that was recently suggested
which is done off-line (no learning involved).  It is based on iteratively
solving the Dynamic Program which arrises out of the optimal control 
problem.  The article is "On the Application of Iterative Dynamic
Programming to Singular Optimal Control Problems," R. Luus, IEEE Trans.
on Autom. Cntrl., Nov. 1992, pp. 1802-1806.  I'm not necessarily suggesting
this method, though.

-Adam Schwartz
adams@robotics.berkeley.edu
U.C. Berkeley