Summary
International Symposium on Nonlinear Theory and its Applications
2010
Session Number:C1L-D
Session:
Number:C1L-D5
Improvement of Tug-of-war Model for Two-armed Bandit Problem:Biologically Inspired Computing Method for Nonlocally-correlated Parallel Searches
Song-Ju Kim, Masashi Aono, Masahiko Hara,
pp.520-523
Publication Date:2010/9/5
Online ISSN:2188-5079
DOI:10.34385/proc.44.C1L-D5
PDF download (165.2KB)
Summary:
The "tug-of-war (TOW) model" proposed in our previous studies [1, 2, 3] is a unique method for parallel searches inspired by the photoavoidance behavior of the single-celled amoeba, the true slime mold Physarum. In the TOW model, many branches of the amoeba act as search agents to collect information on light stimulations while conserving the total sum of their resources (volume). We showed that the nonlocal correlation via resource conservation can be advantageous to manage the "exploration? exploitation dilemma" for solving the multi-armed bandit problem. In this study, we investigate the effect of the information from the other branch on the TOW model’s performance, for the purpose of improving the model. We improve the TOW model so that it can exhibit better performances regardless of the reward probabilities.