C# 클래스 Monopoly.RLHandlers.RLAgent

상속: Monopoly.Classes.Player
파일 보기 프로젝트 열기: pmpailis/RLMonopoly

공개 프로퍼티들

프로퍼티 타입 설명
currentEpoch int
traces List

공개 메소드들

메소드 설명
RLAgent ( ) : System
agent_changeCurrentState ( Observation obs ) : void
agent_cleanup ( ) : void
agent_end ( double reward ) : void
agent_init ( char type, bool policy, string agentName, int inputCount ) : void
agent_start ( Observation observation ) : int
agent_step ( Observation observation, double reward ) : int
createInput ( Observation observation, int action ) : double[]
getNeural ( ) : NeuronDotNet.Core.Network
getRentPayment ( int cp ) : int
getType ( ) : char
initParams ( ) : void
saveOnFile ( string path ) : void
setNeural ( NeuronDotNet.Core.Network net ) : void

비공개 메소드들

메소드 설명
Qlearning ( Observation p_lastState, Monopoly p_lastAction, Observation newState, Monopoly bestAction, double reward ) : double
Sarsa ( Observation lastState, Monopoly lastAction, Observation newState, Monopoly newAction, double reward ) : double
calculateQValues ( Observation obs ) : double[]
checkStateSimilarity ( Observation obs1, Observation obs2 ) : bool
e_greedySelection ( double QValues ) : int
findMaxValues ( double tempQ ) : int
randomAction ( ) : int
trainNeural ( double input, double output ) : void
updateQTraces ( Observation obs, Monopoly a, double reward ) : bool
updateSTraces ( Observation obs, Monopoly a ) : bool

메소드 상세

RLAgent() 공개 메소드

public RLAgent ( ) : System
리턴 System

agent_changeCurrentState() 공개 메소드

public agent_changeCurrentState ( Observation obs ) : void
obs Monopoly.RLClasses.Observation
리턴 void

agent_cleanup() 공개 메소드

public agent_cleanup ( ) : void
리턴 void

agent_end() 공개 메소드

public agent_end ( double reward ) : void
reward double
리턴 void

agent_init() 공개 메소드

public agent_init ( char type, bool policy, string agentName, int inputCount ) : void
type char
policy bool
agentName string
inputCount int
리턴 void

agent_start() 공개 메소드

public agent_start ( Observation observation ) : int
observation Monopoly.RLClasses.Observation
리턴 int

agent_step() 공개 메소드

public agent_step ( Observation observation, double reward ) : int
observation Monopoly.RLClasses.Observation
reward double
리턴 int

createInput() 공개 메소드

public createInput ( Observation observation, int action ) : double[]
observation Monopoly.RLClasses.Observation
action int
리턴 double[]

getNeural() 공개 메소드

public getNeural ( ) : NeuronDotNet.Core.Network
리턴 NeuronDotNet.Core.Network

getRentPayment() 공개 메소드

public getRentPayment ( int cp ) : int
cp int
리턴 int

getType() 공개 메소드

public getType ( ) : char
리턴 char

initParams() 공개 메소드

public initParams ( ) : void
리턴 void

saveOnFile() 공개 메소드

public saveOnFile ( string path ) : void
path string
리턴 void

setNeural() 공개 메소드

public setNeural ( NeuronDotNet.Core.Network net ) : void
net NeuronDotNet.Core.Network
리턴 void

프로퍼티 상세

currentEpoch 공개적으로 프로퍼티

public int currentEpoch
리턴 int

traces 공개적으로 프로퍼티

public List traces
리턴 List