Examples of Reward_observation_terminal

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
        return o;
    }

    public Reward_observation_terminal env_step(Action action) {
        TestUtility.clean_abstract_type(o);
        Reward_observation_terminal ro = null;
        /* Gabor has made it so this environment will step past terminal.  This is
        not something we want to do in general at all.

        But, in order to keep the other tests all working, I'll allow it*/
        if (5 > stepCount) {

            TestUtility.set_k_ints_in_abstract_type(o, 1);
            o.intArray[0] = stepCount;
            stepCount++;

            boolean terminal = stepCount == 5;
            ro = new Reward_observation_terminal(1.0d, o, terminal);
        } else {
            TestUtility.set_k_ints_in_abstract_type(o, 5);
            TestUtility.set_k_doubles_in_abstract_type(o, 5);
            TestUtility.set_k_chars_in_abstract_type(o, 5);


            o.intArray[0] = 173;
            o.intArray[1] = -173;
            o.intArray[2] = 2147483647;
            o.intArray[3] = 0;
            o.intArray[4] = -2147483648;

            o.doubleArray[0] = 0.0078125;
            o.doubleArray[1] = -0.0078125;
            o.doubleArray[2] = 0.0;
            o.doubleArray[3] = 0.0078125e150;
            o.doubleArray[4] = -0.0078125e150;

            o.charArray[0] = 'g';
            o.charArray[1] = 'F';
            o.charArray[2] = '?';
            o.charArray[3] = ' ';
            o.charArray[4] = '&';

            ro = new Reward_observation_terminal(-2.0d, o, false);
        }
        return ro;
    }
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
        return o;  
    }

    public Reward_observation_terminal env_step(Action action) {
        stepCount++;
        Reward_observation_terminal ro=null;
       
        TestUtility.clean_abstract_type(o);
       
        //Short episode with big observations
        if(whichEpisode%2==0){
            TestUtility.set_k_ints_in_abstract_type(o, 50000);
            TestUtility.set_k_doubles_in_abstract_type(o, 50000);

            int terminal=0;
            if(stepCount==200)terminal=1;
                ro=new Reward_observation_terminal(1.0d, o, terminal);
        }
        //Longer episode with smaller obserations
        if(whichEpisode%2==1){
            TestUtility.set_k_ints_in_abstract_type(o, 5);
            TestUtility.set_k_doubles_in_abstract_type(o, 5);

            int terminal=0;
            if(stepCount==5000)terminal=1;
                ro=new Reward_observation_terminal(1.0d, o, terminal);
        }
       
               
        return ro;
    }
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

65
66
67
68
69
70
71
72
    }

    public Reward_observation_terminal env_step(Action action) {
        TestUtility.clean_abstract_type(o);
        int terminal=0;
        Reward_observation_terminal ro=new Reward_observation_terminal(0.0d, o, terminal);
        return ro;
    }
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

64
65
66
67
68
69
70
71
72
73
74
        network.putObservation(obs);
    }

    protected void onEnvStep() {
        Action action = network.getAction();
        Reward_observation_terminal rewardObservation = env.env_step(action);

        network.clearSendBuffer();
        network.putInt(Network.kEnvStep);
        network.putInt(Network.sizeOf(rewardObservation));
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

256
257
258
259
260
261
262
263
264
265
266
267
268
    public static Reward_observation_terminal RL_env_step(Action theAction) {
        checkInstance();
        if (!inited) {
            System.err.println("-- Warning From RLGlue :: RL_env_step() was called without RL_init().");
        }
        Reward_observation_terminal stepResponse = instance.RL_env_step(theAction);
        if (stepResponse == null) {
            System.err.println("-- Warning From RLGlue :: RL_env_step() response was NULL, that should be impossible.");
            stepResponse = new Reward_observation_terminal();
        }
        currentEpisodeOver = stepResponse.isTerminal();
        return stepResponse;
    }
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

70
71
72
73
74
75
76
77
78
79
80
  return nonEmptyObservation;
    }

    public Reward_observation_terminal env_step(Action action) {
       
        Reward_observation_terminal ro=new Reward_observation_terminal();
       
  if(whichEpisode%2==0)
            ro.o=emptyObservation;
        else
            ro.o=nonEmptyObservation;
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
        }
        return theAction;
    }
   
    public synchronized Reward_observation_terminal RL_env_step(Action theAction) {
        Reward_observation_terminal RO = E.env_step(theAction);
        if (RO == null) {
            System.err.println("RO came back as null from RL_step");
        }
        if (RO.getObservation() == null) {
            System.err.println("Ro.o came back as null from RL_step");
        }

        totalReward += RO.getReward();

        if (RO.isTerminal()) {
            numEpisodes++;
        } else {
            numSteps++;
        }
        return RO;
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
    public synchronized Reward_observation_action_terminal RL_step() {
        if (lastAction == null) {
            System.err.println("lastAction came back as null from RL_step");
        }
        Reward_observation_terminal RO=RL_env_step(lastAction);
      

        if (RO.isTerminal()) {
            RL_agent_end(RO.getReward());
        } else {
           lastAction = RL_agent_step(RO.getReward(), RO.getObservation());
        }
        return new Reward_observation_action_terminal(RO.getReward(), RO.getObservation(), lastAction, RO.isTerminal());
    }
View Full Code Here

Examples of org.rlcommunity.rlglue.codec.types.Reward_observation_terminal

131
132
133
134
135
136
137
138
139
140
141
        return obs;
    }

    public synchronized Reward_observation_terminal RL_env_step(Action theAction) {
        send_abstract_type(theAction, Network.kRLEnvStep, "RL_env_step");
        Reward_observation_terminal rot = new Reward_observation_terminal();
        rot.terminal = network.getInt();
        rot.r = network.getDouble();
        rot.o = network.getObservation();
        return rot;
    }
View Full Code Here
TOP
Copyright © 2018 www.massapi.com. All rights reserved.
All source code are property of their respective owners. Java is a trademark of Sun Microsystems, Inc and owned by ORACLE Inc. Contact coftware#gmail.com.