Customization of Algorithm for Progressive Linear Programs (PROLP) Using Neural Net Work

Abstract

In chemical processes, we often take advantage of customizing daily problem-solving when a final solution will be obtained after solving a family of optimization problems repeatedly. In aid of the neural network, in this note we have engaged in improving and customizing the algorithm of progressive linear programs (PROLP) after introducing manifold criteria to estimate active constraints. Through a few numerical experiments, the effect of the approach has been examined.