Feedback Control Based Optimization of Batch Processes in the Reduced Space

The feedback control based optimization of batch processes is to enforce the necessary conditions of optimality in a feedback control manner, instead of explicit numerical optimizations. The control-based optimization does not involve parameter estimation and expensive computations for optimization, while the scheme is able to cope with uncertainties and restore optimality. However, the dimension of necessary conditions of optimality is generally large owing to the input parameterization, which significantly complicates the control task. This paper proposes an approach that can simplify the optimization problem by identifying the principal subspace of plant inputs, which are laid in a small dimension while still retaining most of the optimization performance. A batch reactor is studied to show the effectiveness of the new approach.