Research of large-scale sparse rSQP algorithm based on line search of filter method

In process system optimization, it is common that problem has large number of equality constraints and bound constraints, and relatively few degrees of freedom. Reduced space algorithms are well suited for this category of problem. It is of great importance that sparse structure of specific problem be utilized to save memory and computation time. In this paper a sparse rSQP (reduced sequential quadratic programming) algorithm based on line search of filter method was presented for large-scale process system optimization. This algorithm solves a sequence of QP problems, which are constructed in reduced space by space decomposition and are relatively small in dimensions. Coordination decomposition was used to keep sparse structure of problems, and line search of filter method is incorporated to enhance efficiency. Numerical results in optimizing an example with variable dimension and a distillation column optimization problem demonstrate that the proposed algorithm is quite more efficient than the standard SQP algorithm.