The purpose of this paper is to provide the linearization technique to solve the multidimensional control optimization problem (MCOP) involving first-order partial differential equation (PDEs) constraints. Firstly, we...
详细信息
The purpose of this paper is to provide the linearization technique to solve the multidimensional control optimization problem (MCOP) involving first-order partial differential equation (PDEs) constraints. Firstly, we use the modified objective function approach for simplifying the aforesaid extremum problem (MCOP) and show that the solution sets of the original controloptimizationproblem and its modified controloptimizationproblem (MCOP) Omega$$ {}_{\Omega} $$ are equivalent under convexity assumptions. Further, we use the absolute value exact penalty function method to transform (MCOP) Omega$$ {}_{\Omega} $$ into a penalized controlproblem (MCOP) Omega & rhov;$$ {}_{\Omega \varrho } $$. Then, we establish the equivalence between a minimizer of the modified penalized optimizationproblem (MCOP) Omega & rhov;$$ {}_{\Omega \varrho } $$ and a saddle point of the Lagrangian defined for the modified optimizationproblem (MCOP) Omega$$ {}_{\Omega} $$ under appropriate convexity hypotheses. Moreover, the results established in the paper are illustrated by some examples of MCOPs involving first-order PDEs constraints.
暂无评论