Optimization of a Batch Reactor using NLP and MINLP

This contribution describes the development of nonlinear programming (NLP) and mixed-integer nonlinear programming (MINLP) models for the dynamic optimization of batch reactors. In order to increase the robustness of these models, Orthogonal Collocation on fixed rather than flexible finite elements is applied using Legendre polynomials to explicitly and continuously represent optimal outlet conditions within each element. The NLP model exhibits better robustness and smaller CPU times than the MINLP.