Improved OA Algorithms And Application In Scheduling For Production Process Of Rare Earth

Rare earth is called "industrial vitamins", and it is an important strategic resource. Rare earth reservoirs are very unevenly distributed around the world, and its products and production attract widespread attention. Mathematical model of scheduling problem for rare earth production process is the MINLP model, which contains a lot of continuous, discrete variables and linear and non-linear function. Among the MINLP algorithms, Outer Approximation (OA) algorithm is a common algorithm, but its solution efficiency needs to be improved.The main work of this paper is to further improve the OA algorithm, and apply the presented methods to solve real MINLP cases. Those methods are implemented to solve some benchmark problems and done comparative analysis with standard OA algorithm and the multi-generation cut OA algorithm. We use improved OA algorithms to solve the real cycle scheduling model of ethylene, production scheduling model for cracking furnace and production process scheduling problem for the rare earth. Finally, scheduling optimization system for rare earth production is developed.(1) For the multi-generation cut OA algorithm, two improved algorithms changing the number of parallel sub-problems are proposed:initialization multiple cuts, fixing the number of cut in front and rear for multi-generation cut OA algorithms. Standard numerical tests show that two kinds of improved algorithm can improve the efficiency of the original algorithm in some extent.(2) Adaptive multi-generation cut OA algorithm is presented, which is automatically change the number of cut according to the variations of lower bound. We apply the method in solving the ethylene production scheduling problem. Computational results show that the efficiency of the improved algorithms are significant at iterations and CPU times. With the increasing scale of MINLP problem, there are certain advantages for improved algorithm.(3) Through investigating domestic rare earth enterprises, we refine their features of production process scheduling, and propose rare earth production process scheduling problem, formulating the MINLP optimization model. Using improved OA algorithm to solve it, results show that the improved algorithm can significantly improve the solving efficiency of the practical rare earth production scheduling optimization problem.(4) Based on the investigation of a domestic rare earth plant, we develop a production scheduling optimization system of rare earth. There are functions of real-time production data management and providing scheduling solutions for decision makers. The scheduling scheme can be manually adjusted, and be generated report forms.Improved multi-generation cut OA algorithm is the extension study of effective MINLP algorithms, which has the theoretical significance. The rare earth production process scheduling problem and the mathematical model can improve the efficiency of production process for rare earth, which can provide the basis for scientific management. Corresponding scheduling optimization system is the bridge of practical application and mathematical models.