Presentation 2006-12-14
A Non-linear Approach to Robust Routing Based on Reinforcement Learning
Hideki SATOH,
PDF Download Page PDF download Page Link
Abstract(in Japanese) (See Japanese page)
Abstract(in English) A robust routing algorithm was developed based on reinforcement learning that uses three methods. (1) A state space compression method based on multivariate analysis is used to extract the principal elements from a large number of state variables and compress the state space. (2) A feature space construction method is used to update the orthonormal basis and construct an optimum feature space and to approximate a non-linear function. (3) A search space compression method based on a potential model, a newly developed method, is used to reduce the search space for routing probabilities. This algorithm can take all the network states into account and reduce the adverse effects of disturbance noises. Thus, the frequency of routing loops and falling to a local optimum is reduced. The algorithm works well even if the information it uses is disturbed.
Keyword(in Japanese) (See Japanese page)
Keyword(in English) dynamic routing / robust / reinforcement learning / multivariate analysis / function approximation
Paper # NLP2006-106
Date of Issue

Conference Information
Committee NLP
Conference Date 2006/12/7(1days)
Place (in Japanese) (See Japanese page)
Place (in English)
Topics (in Japanese) (See Japanese page)
Topics (in English)
Chair
Vice Chair
Secretary
Assistant

Paper Information
Registration To Nonlinear Problems (NLP)
Language ENG
Title (in Japanese) (See Japanese page)
Sub Title (in Japanese) (See Japanese page)
Title (in English) A Non-linear Approach to Robust Routing Based on Reinforcement Learning
Sub Title (in English)
Keyword(1) dynamic routing
Keyword(2) robust
Keyword(3) reinforcement learning
Keyword(4) multivariate analysis
Keyword(5) function approximation
1st Author's Name Hideki SATOH
1st Author's Affiliation School of Systems Information Science, Future University-Hakodate()
Date 2006-12-14
Paper # NLP2006-106
Volume (vol) vol.106
Number (no) 414
Page pp.pp.-
#Pages 6
Date of Issue