Skip to content

Struct FAgentState

Struct FAgentState

  • Defined in File AgentState.h

Inheritance Relationships

Base Type

struct FAgentState : public FInitialAgentState

Source: Source/Schola/Training/Public/TrainingDataTypes/AgentState.h

Dependencies: FInitialAgentState, FPoint

Agent state returned after an environment step.

Extends FInitialAgentState with reward, done, and truncated flags following the gym convention.


Public Functions

SymbolDetails
FAgentStateConstructor with full state information.
FAgentStateDefault constructor.

FAgentState

inline FAgentState(const TInstancedStruct<FPoint> &InPoint, float Reward,
bool bTerminated, bool bTruncated,
const TMap<FString, FString> &InInfo)

Constructor with full state information.

Parameters

  • InPoint – [in] The observation.

  • Reward – [in] The reward value.

  • bTerminated – [in] Whether the episode is complete.

  • bTruncated – [in] Whether the episode was truncated.

  • InInfo – [in] The info dictionary.

#DirectionNameTypeDescription
1InPointconst TInstancedStruct< FPoint > &The observation.
2RewardfloatThe reward value.
3bTerminatedboolWhether the episode is complete.
4bTruncatedboolWhether the episode was truncated.
5InInfoconst TMap< FString, FString > &The info dictionary.

Attributes: inline

Source: Source/Schola/Training/Public/TrainingDataTypes/AgentState.h

FAgentState

inline FAgentState()

Default constructor.

#DirectionNameTypeDescription
1InPointconst TInstancedStruct< FPoint > &The observation.
2RewardfloatThe reward value.
3bTerminatedboolWhether the episode is complete.
4bTruncatedboolWhether the episode was truncated.
5InInfoconst TMap< FString, FString > &The info dictionary.

Attributes: inline

Source: Source/Schola/Training/Public/TrainingDataTypes/AgentState.h


Public Members

SymbolDetails
RewardThe scalar reward received after taking the action.
bTerminatedWhether the episode has ended due to task completion.
bTruncatedWhether the episode was truncated (e.g., time limit reached).

Reward

float Reward = 0.0f

The scalar reward received after taking the action.


bTerminated

bool bTerminated = false

Whether the episode has ended due to task completion.


bTruncated

bool bTruncated = false

Whether the episode was truncated (e.g., time limit reached).