Dissertation > Excellent graduate degree dissertation topics show

Fitness Landscapes Analysis for Perturbative Hyper-Heuristics

Author: JiangYi
Tutor: JiangHe
School: Dalian University of Technology
Course: Software Engineering
Keywords: Fitness Landscapes Analysis Constructive Hyper-Heuristics PerturbativeHyper-Heuristics Local Optimal Network Escape Analysis
CLC: TP18
Type: Master's thesis
Year: 2013
Downloads: 19
Quote: 0
Read: Download Dissertation

Abstract


Fitness landscapes analysis has been applied to analyze the structure of constructive hyper-heuristics space, which reveals these landscapes have a high neutrality and positional bias. They also have the feature of a globally convex or big valley structure. Ochoa conducted a fitness landscapes analysis on a hyper-heuristic search space by using graph coloring heuristics for timetabling. The result confirmed that the optimal solution was not isolated in the space, but surrounded by many local optimal solutions. We can use these futures to design more efficient hyper-heuristics. In the other related paper, Ochoa performed a fitness landscapes analysis of the hyper-heuristic space induced by a dispatching-rule-based hyper-heuristic for hybrid flowshop scheduling problem. This study confirmed that it was suitable for hyper-heuristic space by using fitness landscapes analysis and it was related between them.However, there is no study of applying fitness landscapes analysis for perturbative hyper-heuristics space. One reason is that the process of building a solution of a problem instance is different for constructive hyper-heuristics and perturbative hyper-heuristics: constructive hyper-heuristics start with an empty solution, and build a full solution step by step; perturbative hyper-heuristics start with a full solution, and gradually optimize this solution. Constructive hyper-heuristics will generate a certain solution after execute a heuristic algorithm sequence, but for perturbative hyper-heuristics, the solution is uncertain. Moreover, the initial solution has a big influence on the perturbative hyper-heuristics, meanwhile the initial solution has no influence on constructive hyper-heuristics. Based on these reasons, there is no research on perturbative hyper-heuristics space by using fitness landscapes analysis so far. The purpose of this paper is to make up the gap. Some methods of this paper use to reduce the impact of initial solution on perturbative hyper-heuristics, and we seek to use the average of multiple solutions fitness as the fitness of heuristic algorithm sequence. Furthermore, we use the notion of fitness distance correlation to explore the relationship of fitness and distance between local optimal solutions and global optimal solution, and find positive relationship between them; we use escape analysis to find a perfect mutation ratio for local optimal solutions, which can guide the algorithm design.

Related Dissertations

  1. Incremental Data Race Detection,TP311.11
  2. Research on Java’s Synchronization Optimization Based on Escape Analysis,TP311.10
  3. Fault Diagnosis Method Based on Support Vector Machine,TP18
  4. Designs and Applications of Fuzzy Synthetic Evaluation Models Based on Parallel Algorithms,TP18
  5. Public Transport Optimal Dispatching Based on the Genetic-Newton Algorithm,TP18
  6. Research on Clustering Algorithm Based on Mutation Particle Swarm Optimization,TP18
  7. Research on Feature Selection and Construction in Emotion Speech Recognition,TP18
  8. Research on Fuzzy C-Mean Clustering Algorithm Based on Particle Swarm Optimization and Shuffled Frog Leaping Algorithm,TP18
  9. Research on Clustering Algorithm Based on Genetic Algorithm and Rough Set Theory,TP18
  10. Design and Implementation of Cross-selling System Based on Knowledge Reasoning,TP18
  11. Based on Rough Set of Urban Areas When Traffic Green Control System Research,TP18
  12. Research of Power System Reactive Power Optimization Based on Immune Ant Colony Algorithm,TP18
  13. Philosophy of artificial intelligence,TP18
  14. Research on Theory of Granular Computing and Its Application on Image Retrieval,TP18
  15. Research on the Improvements and Applications of Particle Swarm Optimization,TP18
  16. Incremental rough set attribute reduction,TP18
  17. Calculation of Knowledge Granulation and Study of Its Application in Attribute Reduction,TP18
  18. Reduction and Rule Acquisition in Fuzzy Decision System Based on Neighborhood Relationship,TP18
  19. The Application of Plan Recognition in Monitoring the Money Flow for the Stock Market,TP18
  20. Fuzzy Multi-atrribute Decision Making Based on Vague Sets,TP18
  21. Researches on Improved Genetic Algorithm Base on Reinforcement Learning,TP18
  22. Stage Fault Diagnosis Based on Data Fusion,TP18

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