Class ISingleAgentScholaEnvironment

Class ISingleAgentScholaEnvironment

• Defined in File SingleAgentEnvironmentInterface.h

Inheritance Relationships

Base Type

• public IBaseScholaEnvironment (Class IBaseScholaEnvironment)

class ISingleAgentScholaEnvironment : public IBaseScholaEnvironment

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

Dependencies: FAgentState, FInitialAgentState, FInteractionDefinition, FPoint

Interface for single-agent reinforcement learning environments.

Implement this interface in your environment class to create a single-agent RL environment. All methods are Blueprint Native Events, allowing implementation in either C++ or Blueprint.

---

Public Functions

Symbol	Details
InitializeEnvironment	Initialize the environment and define the agent’s observation and action spaces.
Reset	Reset the environment to its initial state.
Step	Execute one environment step with the given action.
SeedEnvironment	Set the random seed for reproducible environment behavior.
SetEnvironmentOptions	Configure the environment with custom options.

InitializeEnvironment

void InitializeEnvironment(FInteractionDefinition &OutAgentDefinition)

Initialize the environment and define the agent’s observation and action spaces.

Parameters

OutAgentDefinition – [out] The interaction definition containing observation/action space information.

#	Direction	Name	Type	Description
1	—	OutAgentDefinition	FInteractionDefinition &	The interaction definition containing observation/action space information.

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

Reset

void Reset(FInitialAgentState &OutAgentState)

Reset the environment to its initial state.

Parameters

OutAgentState – [out] The initial agent state after reset, including the first observation.

#	Direction	Name	Type	Description
1	—	OutAgentState	FInitialAgentState &	The initial agent state after reset, including the first observation.

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

Step

void Step(const FInstancedStruct &InAction, FAgentState &OutAgentState)

Execute one environment step with the given action.

Parameters

• InAction – [in] The action selected by the agent for this step.

• OutAgentState – [out] The resulting agent state (observation, reward, done flags, info).

#	Direction	Name	Type	Description
1	—	InAction	const FInstancedStruct &	The action selected by the agent for this step.
2	—	OutAgentState	FAgentState &	The resulting agent state (observation, reward, done flags, info).

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

SeedEnvironment

void SeedEnvironment(int Seed)

Set the random seed for reproducible environment behavior.

Parameters

Seed – [in] The random seed value.

#	Direction	Name	Type	Description
1	—	Seed	int	The random seed value.

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

SetEnvironmentOptions

void SetEnvironmentOptions(const TMap<FString, FString> &Options)

Configure the environment with custom options.

Parameters

Options – [in] Map of configuration option names to values.

#	Direction	Name	Type	Description
1	—	Options	const TMap< FString, FString > &	Map of configuration option names to values.

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

---

Public Static Functions

Symbol	Details
Execute_InitializeEnvironment	Adapter method to convert single-agent initialization to multi-agent format.
Execute_Reset	Adapter method to convert single-agent reset to multi-agent format.
Execute_Step	Adapter method to convert single-agent step to multi-agent format.

Execute_InitializeEnvironment

static inline void Execute_InitializeEnvironment(
    UObject *Obj, TMap<FString, FInteractionDefinition> &OutAgentDefinitions)

Adapter method to convert single-agent initialization to multi-agent format.

Parameters

• Obj – [in] The environment object.

• OutAgentDefinitions – [out] Map of agent IDs to definitions (will contain one entry “SingleAgent”).

#	Direction	Name	Type	Description
1	—	Obj	UObject *	The environment object.
2	—	OutAgentDefinitions	TMap< FString, FInteractionDefinition > &	Map of agent IDs to definitions (will contain one entry “SingleAgent”).

Attributes: inline, static

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

Execute_Reset

static inline void
Execute_Reset(UObject *Obj, TMap<FString, FInitialAgentState> &OutAgentState)

Adapter method to convert single-agent reset to multi-agent format.

Parameters

• Obj – [in] The environment object.

• OutAgentState – [out] Map of agent IDs to states (will contain one entry “SingleAgent”).

#	Direction	Name	Type	Description
1	—	Obj	UObject *	The environment object.
2	—	OutAgentState	TMap< FString, FInitialAgentState > &	Map of agent IDs to states (will contain one entry “SingleAgent”).

Attributes: inline, static

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h

Execute_Step

static inline void
Execute_Step(UObject *Obj,
             const TMap<FString, TInstancedStruct<FPoint>> &InActions,
             TMap<FString, FAgentState> &OutAgentStates)

Adapter method to convert single-agent step to multi-agent format.

Parameters

• Obj – [in] The environment object.

• InActions – [in] Map of agent IDs to actions (expects one entry “SingleAgent”).

• OutAgentStates – [out] Map of agent IDs to states (will contain one entry “SingleAgent”).

#	Direction	Name	Type	Description
1	—	Obj	UObject *	The environment object.
2	—	InActions	const TMap< FString, TInstancedStruct< FPoint > > &	Map of agent IDs to actions (expects one entry “SingleAgent”).
3	—	OutAgentStates	TMap< FString, FAgentState > &	Map of agent IDs to states (will contain one entry “SingleAgent”).

Attributes: inline, static

Source: Source/Schola/Training/Public/Environment/SingleAgentEnvironmentInterface.h