A Nonlinear Multiperiod Process Optimization Model for Prodution Planning in Multi-plant Facilities

In this paper we propose a multiperiod nonlinear programming (NLP) formulation that incorporates empirical process models for the optimal planning of a multi-plant production site. Using as a basis a real world application of a polymer plant that produces 27 products, a model is developed for predicting the detailed production using actual plant data. The empirical process models account for raw material usage, physical product specifications (e.g. viscosity), operation limitations, and production rates. NLP models are proposed for each of the plants, and a multiperiod NLP model is then formulated that determines monthly production and inventory levels of all products for each plant. The NLP model for the polymer plant has several thousand variables and constraints, and a web interface was developed so many users can access the model over the intranet. A graphical input template is linked to the optimization model in which the user can modify key input variables as well as operating and demand parameters. Several numerical examples are presented to illustrate the scope of this model.