Action Selection methods using Reinforcement Learning

摘要: Action Selection schemes, when translated into precise algorithms, typically involve considerable design effort and tuning of parameters. Little work has been done on solving the problem using learning. This paper compares eight different methods action selection Reinforcement Learning (learning from rewards). The range centralised cooperative to decentralised selfish. They are tested in an artificial world their performance, memory requirements reactiveness compared. Finally, possibility more exotic, ecosystem-like models considered.