Introduction to Reinforcement Learning
Introduction to Reinforcement Learning
| Hi-index | 0.01 |
As the casts of characters in computer games grow bigger, there is a necessity to move away from hand coded behaviours and to employ learning techniques. This paper introduces a new approach to reinforcement learning for large multi-agent systems as found in massively multi-player computer games. It is proposed that, an agent can be assigned one or more action-selection policies simultaneously, policies can be assigned partially, policies can be shared among agents and, agents can switch between policies at any time. Game conductors are used for controlling the plot of a game. An implementation of simple emotions incites changes in the behaviour of characters.