Significance elf problem structure in chemical process optimization strategies

Mathematical programming techniques have been used extensively to solve chemical engineering problems, covering activities from process design and synthesis, to real-time optimization, planning and scheduling. The corresponding mathematical representations range from linear/nonlinear programming (LP/NLP) to mixed integer linear/nonlinear programming (MILP/MINLP). Large scale optimization of complex systems requires full utilization of the particular structure, in order to achieve solution robustness and efficiency. This issue can be addressed at two levels: (i) at the strategic level of designing a tailored solution method, and (ii) at the algorithmic level of exploiting the structure of the Jacobian matrix in the sparse matrix computations. The latter is particularly important since it lies in the foundation of all optimization methodologies. The above issues are illustrated through applications to chemical engineering design problems.