Research Article
Performance Evaluation of Multiagent Reinforcement Learning Based Training Methods for Swarm Fighting
| Input: Begin//bool value, initialize the environment, | | whether to start round. | | Output: Winner of this round. | | Execute this program each frame. | | MaxEnviormentStep =8000//Defines the maximum. | | number of steps in the environment. | | for MaxEnviormentStep && Begin do. | | ++ | | if Red.num ==0 && Green.num ==0 then | | return tie | | else if Red.num ==0 then | | return Green.win | | else | | return Red.win | | return null |
|
