Exploration and Exploitation Trade-Off in Multiagent Learning
Keiki Takadama, Katsunori Shimohara · Computational intelligence · 2001
This paper focuses on the trade-off between exploration and exploitation in multiagentlearning and explores some fundamental factors that contribute to clarifying this trade-offThrough inventive simulations on distributed constraint satisfaction problems in multiagentenvironments, the following implications are revealed: (1) the trade-off between explorationand exploitation at the collective level is not easy to be solved when considering the trade-offat the individual level; but (2) the trade-off at the collective level can be solved by introducinga simple gradient search in the trade-off at the individual level.