IPortfolioAgent<T>

Interfaces · AiDotNet.Finance.Trading.Evaluation

The MINIMAL agent surface the portfolio trainer needs: choose an action, remember a transition, learn, and reset per episode. Kept narrow (rather than the full IRLAgent, which pulls in the entire IFullModel surface) so the trainer depends on nothing more than the RL loop, and a test double is trivial. Wrap any real AiDotNet agent with From.

Known implementations: FeedforwardPolicyAgent<T>, RecurrentPolicyAgent<T>

Methods

Method Summary
T Train() Runs one learning update from the stored experience.
Vector<T> SelectAction(Vector<T> state, bool explore) Chooses the portfolio action for state.
void ResetEpisode() Clears the agent’s per-episode state before the next episode starts.
void StoreExperience(Vector<T> state, Vector<T> action, T reward, Vector<T> nextState, bool done) Records one transition for the agent to learn from.