burlap episode

burlap episode

Each episode consists of the agent moving from the initial state to the goal state

//run learning for 50 episodes
 for(int i = 0; i < 50; i++){
 Episode e = agent.runLearningEpisode(env);

 e.write(outputPath + "ql_" + i);
 System.out.println(i + ": " + e.maxTimeStep());

 //reset environment for next learning episode
 env.resetEnvironment();
 }

发表评论

电子邮件地址不会被公开。 必填项已用*标注