Dissertation > Excellent graduate degree dissertation topics show

Research on Application of Bounded Derivative Neural Network for Nonlinear Predictive Control

Author: ShiYiPing
Tutor: ZhaoJun
School: Zhejiang University
Course: Systems Engineering
Keywords: Bounded derivative neural network Predictive Control Gain scheduling
CLC: TP183
Type: Master's thesis
Year: 2008
Downloads: 55
Quote: 2
Read: Download Dissertation


The learning process of the neural network is an optimization process , depending on the error information to reasonably select the network weights . In the input and output of the network model of the relationship established , the most important guiding information is a derivative of the relationship between the input and output , this derivative relations can only properly establish the corresponding relationship between input and output values ??. Currently, the optimization of the neural network learning algorithm , using only the network output data error as guidance information for network training , and no effective way of bringing a derivative of the model , lead to poor network generalization ability , not practical . Simple minimum output data error as standard to guide network is trained , its essence is only accurate interpolation of the sample data , and the non - smooth interpolation to generate a large error , thus the non - sample data . Therefore , the network training process , in addition to considering the information provided by the sample data , should also take into account the derivative of the model , and turn it into a form of constraint is introduced into the the network learning optimization process which . This paper studies the neural network training algorithms with constraint , and neural network applications to nonlinear model predictive control . The main content of this paper and innovations include : (1) using two different methods with constrained training of the neural network , through penalty function neural network and bounded derivative of the neural network training results , the analysis of these two methods exist advantages and disadvantages. (2 ) using the bounded derivative neural network \\ modeling of the steady-state model , combined with the DMC algorithm , real-time forecast model gain scheduling , to some extent, improve the quality of control . (3) by changing the bounded derivative of neural networks of the network structure , the network from the static network changes for dynamic network , while still making the network maintained its static has the derivative constraints capacity , so that the bounded derivative neural network has a the modeling capability can be the object of the static and dynamic two-part .

Related Dissertations

  1. Satellite Attitude Control Methods Using Only Magnetorquers,V448.222
  2. Research of Predictive Control in Water Treatment Plant Based on Neuron Network,TP273.1
  3. Robust Model Predictive Control of Networked Systems,TP273
  4. Control System Design of Electroslag Continuous Caster and Control Scheme Study on Casting Speed,TF345
  5. Air-cooled proton exchange membrane fuel cell dynamics control and performance testing of,TM911.4
  6. Distributed Coordination iron production process optimization methods,TF58
  7. Research on a Class of Uncertain Dmc Method in Blast Furnace,TF543
  8. Nonlinear Predictive Control for Polypropylene Grade Transition Process Based on Neural Network Modeling,TQ325.14
  9. Temperature control and design of proton exchange membrane fuel cell,TM911.4
  10. Predictive Control for EAF Electrode Regulating System with Dead-zone,TF741.5
  11. Based on the study of the boiler water level of the PSO algorithm generalized predictive control,TP273
  12. Research on Generalized Predictive Control Algorithm in Ship Position of Replenishment Ship,U661.33
  13. Research on Model Predictive Control of Dynamic Positioning System for Crane Vessels,U674.35
  14. Based on the research and development of the ARM model predictive control system,TP273
  15. 1000MW Ultra Supercritical Units in the Main Steam Temperature Control System’s Characteristics and Control Strategy Research,TM621.6
  16. Strene Device Control System Optimization and Implementation,TP273
  17. Aromatics distillation unit RMPCT advanced control system design and implementation,TP273
  18. Research on Steady and Dynamic Behaviors of a Novel Heat Integration System with Divided Wall Column and Double-effect Distillation,TQ028.13
  19. Ferromanganese smelting electric arc furnace electrode control system simulation,TP273
  20. Fuzzy Modelling and Control of Wastewater Treatment Processes,X703
  21. Study on Networked Predictive Control for Serially Connected Systems,O231

CLC: > Industrial Technology > Automation technology,computer technology > Automated basic theory > Artificial intelligence theory > Artificial Neural Networks and Computing
© 2012 www.DissertationTopic.Net  Mobile