Some algebraic properties of optimization problems in complex chemical plants

Abstract The determination of optimum conditions in a chemical plant comprising a number of interconnected units often presents considerable computational difficulties because of the large number of parameters which must be simultaneously varied. The method of dynamic programming permits the problem to be decomposed into a set of sub-problems of lower dimensionality, but is limited in application to systems consisting of simple sequential chains of units. The present work describes a classical variational approach which permits a similar dimensional decomposition to be effected in plants of arbitrarily complex structure. A number of systems which exemplify the main features of the method without undue algebraic complexity are discussed in detail.