Tug-of-war Model for Competitive Multi-armed Bandit Problem: Amoeba-inspired Algorithm for Cognitive Medium Access

Kim Song-Ju, Aono Masashi, Nameda Etsushi, Hara Masahiko

doi:10.15248/proc.1.590

The “tug-of-war (TOW) model” is a unique parallel search algorithm for solving the multi-armed bandit problem (BP), which was inspired by the photoavoidance behavior of a single-celled amoeboid organism, the true slime mold Physarum polycephalum [1, 2, 3, 4, 5, 6]. “The cognitive medium access”, which refers to multiuser channel allocations in cognitive radio, can be interpreted as “competitive multi-armed bandit problem (CBP) [14].” Unlike the normal BP, the reward (free channel) probability of a channel selected by more than one user is evenly split between selecting users. In this study, we propose the “solid TOW (STOW) model” for the CBP toward developing cognitive medium access protocols in uncertain environments. The aim of this study is to explore how can the users achieve the “social maximum”, which is the most desirable state to obtain the maximum total score, in a decentralized manner. We show that the performance of the STOW model is higher than that of the well-known UCB1-tuned algorithm in many cases.

Tug-of-war Model for Competitive Multi-armed Bandit Problem: Amoeba-inspired Algorithm for Cognitive Medium Access

Description

Journal

Citations (2)*help

References(10)*help

Keywords

Details 詳細情報について

Export

Report a problem