本文整理汇总了Java中burlap.behavior.singleagent.planning.stochastic.sparsesampling.SparseSampling.setForgetPreviousPlanResults方法的典型用法代码示例。如果您正苦于以下问题：Java SparseSampling.setForgetPreviousPlanResults方法的具体用法？Java SparseSampling.setForgetPreviousPlanResults怎么用？Java SparseSampling.setForgetPreviousPlanResults使用的例子？那么, 这里精选的方法代码示例或许可以为您提供帮助。您也可以进一步了解该方法所在类burlap.behavior.singleagent.planning.stochastic.sparsesampling.SparseSampling的用法示例。
在下文中一共展示了SparseSampling.setForgetPreviousPlanResults方法的2个代码示例，这些例子默认根据受欢迎程度排序。您可以为喜欢或者感觉有用的代码点赞，您的评价将有助于系统推荐出更棒的Java代码示例。
示例1: IPSS
import burlap.behavior.singleagent.planning.stochastic.sparsesampling.SparseSampling; //导入方法依赖的package包/类
public static void IPSS(){
InvertedPendulum ip = new InvertedPendulum();
ip.physParams.actionNoise = 0.;
Domain domain = ip.generateDomain();
RewardFunction rf = new InvertedPendulum.InvertedPendulumRewardFunction(Math.PI/8.);
TerminalFunction tf = new InvertedPendulum.InvertedPendulumTerminalFunction(Math.PI/8.);
State initialState = InvertedPendulum.getInitialState(domain);
SparseSampling ss = new SparseSampling(domain, rf, tf, 1, new SimpleHashableStateFactory(), 10 ,1);
ss.setForgetPreviousPlanResults(true);
ss.toggleDebugPrinting(false);
Policy p = new GreedyQPolicy(ss);
EpisodeAnalysis ea = p.evaluateBehavior(initialState, rf, tf, 500);
System.out.println("Num steps: " + ea.maxTimeStep());
Visualizer v = InvertedPendulumVisualizer.getInvertedPendulumVisualizer();
new EpisodeSequenceVisualizer(v, domain, Arrays.asList(ea));
}
示例2: IPSS
import burlap.behavior.singleagent.planning.stochastic.sparsesampling.SparseSampling; //导入方法依赖的package包/类
public static void IPSS(){
InvertedPendulum ip = new InvertedPendulum();
ip.physParams.actionNoise = 0.;
RewardFunction rf = new InvertedPendulum.InvertedPendulumRewardFunction(Math.PI/8.);
TerminalFunction tf = new InvertedPendulum.InvertedPendulumTerminalFunction(Math.PI/8.);
ip.setRf(rf);
ip.setTf(tf);
SADomain domain = ip.generateDomain();
State initialState = new InvertedPendulumState();
SparseSampling ss = new SparseSampling(domain, 1, new SimpleHashableStateFactory(), 10, 1);
ss.setForgetPreviousPlanResults(true);
ss.toggleDebugPrinting(false);
Policy p = new GreedyQPolicy(ss);
Episode e = PolicyUtils.rollout(p, initialState, domain.getModel(), 500);
System.out.println("Num steps: " + e.maxTimeStep());
Visualizer v = CartPoleVisualizer.getCartPoleVisualizer();
new EpisodeSequenceVisualizer(v, domain, Arrays.asList(e));
}