HANDLING IMPLICIT MODEL FORMULATIONS IN MINLP OPTIMIZATION

This paper deals with the handling of implicit model formulations in M INLP optimization and proposes a strategy to cope with occurring diffi culties of insufficient information. Process flowsheet simulators are the state-of-the-art tool of chemical engineers. Since these packages often utilize implicit model formulations for physical properties calc ulation etc., it is necessary to analyze the performance of MINLP algo rithms when dealing with such problem representations. The examination s are carried out with the program system PSANO (Process Synthesis and Network Optimization), where the equation-oriented flowsheeting packa ge SPEEDUP and the Mixed Integer Linear Programming (MILP) solver OSL have been integrated into an extension of the Outer Approximation (OA) approach.