net.pakl.rl
Class AgentParallelized

java.lang.Object
  net.pakl.rl.Agent
      net.pakl.rl.AgentParallelized

public class AgentParallelized
extends Agent

This class represents a multi-threaded agent (therefore faster if you have multiple CPUs) appropriate for reinforcement learning with full Value Iteration, since during value iteration each state can be updated independently of others (as long as testTrainSameVf is false).

Field Summary

protected int

threadsAtATime

Fields inherited from class net.pakl.rl.Agent

callNumber, DEFAULT_RANDOM_NUMBER_SEED, discountFactor, epsilon, greed, name, policy, printStartStateValueEvery, random, reinforcementFunction, state, UPDATE_ONLY_WHEN_GREEDY, world

Constructor Summary

AgentParallelized(int numThreads)

AgentParallelized(java.lang.String newName, int numThreads)

Method Summary

ValueFunction

performValueIteration(ValueFunction newValueFunction, ValueFunction valueFunction) This is the MAIN learning function in the Agent.

Methods inherited from class net.pakl.rl.Agent

experience, getAverageDelta, getBestActionForValueFrom, getDiscountFactor, getMaximumDelta, getName, getTotalDelta, initializeRandomSeed, performValueIterationTrajectorySample, performValueIterationUpdateOnState, setDiscountFactor, setEpsilon, setGreed, setPolicy, setReinforcementFunction, setState, setWorld

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Field Detail

threadsAtATime

protected int threadsAtATime

Constructor Detail

AgentParallelized

public AgentParallelized(java.lang.String newName,
                         int numThreads)

AgentParallelized

public AgentParallelized(int numThreads)

Method Detail

performValueIteration

public ValueFunction performValueIteration(ValueFunction newValueFunction,
                                           ValueFunction valueFunction)

Description copied from class: Agent

This is the MAIN learning function in the Agent.

Overrides:

performValueIteration in class Agent