Research On TLBO Algorithm Based Modeling Methods For Chemical Processes

Teaching-learning based optimization(TLBO)is a new kind of intelligent algorithm that simulates the teaching behavior of the teacher,which has the characteristics of fast convergence and simple implementation.This thesis focus on the improvements of TLBO and its applications in chemical process modeling.The main contents of this thesis are as follows:1)This thesis briefly a review on the development of TLBO,and the chemical pro-cess modeling methods.2)The hybrid particle teaching-learning based optimization(HPTLBO)is proposed.Particle swarm optimization(PSO)has the advantages of short searching time and higher searching precision.In HPTLBO,PSO with cross-operation and variable weights is com-bined with TLBO adding extra-curricular learning phase,and the mutation operation is added.Some benchmark functions are used to verify the effectiveness of HPTLBO.HPT-LBO is applied to optimize the parameters of support vector machine for regression(SVR),then the SVR model of the distillations of WB and OR column is established.The simulation results show the effectiveness of the HPTLBO-SVR model.3)The modified TLBO inspired by fish swarm behavior(MFTLBO)is proposed.The crowding factor in artificial fish algorithm(AF)is used to define the teacher's control binding of some students,who is difficult to restrict.Using chaos search in MFTLBO when the step before doesn't work well.By comparing MFTLBO algorithm with some intelligent optimization algorithms for some benchmark functions,it can be seen that the performance of MFTLBO is better.MFTLBO is used to optimize the parameters of the Radial Basis Function neural network(RBF-NN)for modeling a FCCU reaction-regen-eration process.The simulation results show the effectiveness of the MFTLBO-RBF model.