2014 Fiscal Year Final Research Report
Parameter Decision Method of Chaotic Search for Combinatorial Optimization Problem
| Project/Area Number |
25870770
|
|---|
| Research Category |
Grant-in-Aid for Young Scientists (B)
|
| Allocation Type | Multi-year Fund |
|---|
| Research Field |
Soft computing
Social systems engineering/Safety system
|
|---|
| Research Institution | Nippon Institute of Technology (2014)
Tokyo University of Science (2013) |
|---|
Principal Investigator |
|
|---|
| Project Period (FY) |
2013-04-01 – 2015-03-31
|
| Keywords | 組合せ最適化問題 / 発見的解法 / カオスニューラルネットワーク / ソフトコンピューティング |
|---|
| Outline of Final Research Achievements |
For finding near optimal solutions of NP-hard combinatorial optimization problems, a chaotic search method has been proposed. In the method, a chaotic neural network is constructed by chaotic neurons. We have already clarified that its high searching ability depends on a statistical property of refractory effects in the chaotic neuron. However it is difficult to find optimal parameters of the refractory effect. In order to develop a method that decides optimal parameters, we confirmed relation between the values of parameters and searching abilities of the chaotic search method or a dynamics of the chaotic neuron from some numerical experiments.
We analyzed relation the searching ability and Lyapunov exponent of a single chaotic neuron. A positive Lyapunov exponent indicates that a behavior of the chaotic neuron is chaos. From some numerical experiments, we confirmed that if the Lyapunov exponents are positive, good results are obtained for a motif extraction problem.
|
| Free Research Field |
数理情報学
|
