Dissertation > Excellent graduate degree dissertation topics show

Multi-directional Mutation Genetic Algorithm and Research on Neural Network Optimization

Author: LiZhenYe
Tutor: HuJinSong;TianWenChun
School: South China University of Technology
Course: Computer technology
Keywords: Genetic Algorithms BP network optimization algorithms multi-directional
CLC: TP18
Type: Master's thesis
Year: 2011
Downloads: 27
Quote: 0
Read: Download Dissertation

Abstract


Many aspects in real life related to the optimization problems such as path planning, network optimization, economic investment and so on. Optimization problems are present in almost all areas of the real word. Most of the optimization problems are actually some complex high-dimensional nonlinear coupling function, and the expressions of these functions are usually unknown. Using the traditional mathematical methods for example Linear programming to solve thease problems can not get a satifactory result.In recent decades, along with the rapid development of computational intelligence technology, some intelligent optimization algorithms that are different from the classic mathematical methods have appeared. Among them, the Genetic Algorithm with its excellent global search ability, implicit parallelism and simple operation operator received academic acclaim, and thus becomes one of the hottest technologies to solve the optimization problems. BP neural network is an error back propagation artificial neural network that trained by the multilayer feedforward based on gradient descent algorithm. Its calculation is simple, easy, small and with the characteristics of strong parallelism. The Genetic Algorithm and BP neural network can complement with each other.However, after a large number literature reading and function optimization experiments, in particular, genetic algorithm and BP neural network experiments, we found that it is inadequate for genetic algorithm to solve some of high-dimensional, non-linear separation function. It’s still likely to fall into local minima and the optimization efficient is low. For these reasons, this paper presents a multi-directional mutation genetic algorithm. By mixing a variety of encoding methods, genetic algorithm can overcome the disadvantage of population’s "cross" distribution. And by introducing a new population, the GA enhances the local search ability.Through a series of experiments, this paper compares the multi-directional mutation genetic algorithm to the simple genetic algorithm, the general two-population genetic algorithm and PSO. We found that our method is more efficient and better.Finally, we use the proposed multi-directional genetic algorithm to solve the BP network problem and make a good result.

Related Dissertations

  1. Development of the on-line Training and Examination System of Army,TP311.52
  2. Designs and Applications of Fuzzy Synthetic Evaluation Models Based on Parallel Algorithms,TP18
  3. BP network optimization based on genetic algorithm optimization of the biodiesel process,TE667
  4. Based on Genetic Algorithm Pishihang irrigation canal water allocation marshalling model of,S274
  5. Genetic Algorithm in logistics and warehousing Optimization Research,F259.2
  6. Mining resources based on genetic algorithm optimization model of,O224
  7. Digital image forensics technology research,TP391.41
  8. The Research and Application of Modified Algorithms About Fuzzy Predictive Functional Control,TP273
  9. Optimal Control of Emulsion System in Cold Rolling,TP273
  10. Research on the Marshalling-scheduling Model and Algorithms of Freight Trains Based on Game Theory,O225
  11. The Application of Using Genetic Algorithms on Universities Course-arranging System,TP18
  12. Research on Mobile Robot Path Planning and Simulation Realization,TP242
  13. Research on Routing Algorithmin Sensor Networks Based on Cluster with Mobile Sink,TP212.9
  14. Research and Implement of the Theme Crawler for Automotive Industry,TP391.3
  15. The Research and Implement on Camera Calibration Technology Based on Trifocal Tensor,TP391.41
  16. Decision Support System of Vehicle Scheduling in Double Level Garage,TP242
  17. The Studies on Some Improvements of the GA and Their Applications in SVM,TP18
  18. A Reduction Method for Artificial Neural Network Inputs Based on An Improved Genetic Algorithm,TP18
  19. Optimal Riser Design of Steel Casting Based on CAE Analysis,TG260
  20. In-furnace Temperature Information Included Combustion Optimization of a Utility Boiler,TK227.1

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