Вычисление объединенных мягких ограничений в задачах проектирования оптимальных технологических систем

Abstract

Taking into account the uncertainty in the initial information in the formulation of chemical process optimization problems leads to the use of various forms of constraints representing the design requirements for the operation of the chemical process. The paper proposes a method to solve optimal chemical process design problems, taking into account joint chance probabilistic constraints. The method is based on the use of piecewise linear approximation of the constraint functions in left parts of the joint constraints as well as objective function. Approximation of the constraints satisfaction region in the form of multidimensional parallelepipeds is used. This allows us to reduce the joint chance constraints to a set of separate deterministic constraints and provide their computability, as well as less time to get a problem solution. The used approximations are made more exact by uncertainty region and the multidimensional parallelepipeds partitioning.