- Conference Paper
In this paper, we propose a novel formulation for encoding state constraints into the Linear Programming approach to Approximate Dynamic Programming via the use of penalty functions. To maintain tractability of the resulting optimization problem that needs to be solved, we suggest a penalty function that is constructed as a point-wise maximum taken over a family of low-order polynomials. Once the penalty functions are designed, no additional approximations are introduced by the proposed formulation. The effectiveness and numerical stability of the formulation is demonstrated through examples Show more
Book title20th IFAC World Congress. Proceedings
Journal / seriesIFAC-PapersOnLine
Pages / Article No.
SubjectStochastic optimal control problems; Approximate dynamic programming; Soft constraints
Organisational unit03751 - Lygeros, John / Lygeros, John
MoreShow all metadata