Struct openrailwaymap_exporter::ai::TrainAgentAI

pub struct TrainAgentAI {
    pub id: RailwayObjectId,
    pub railway_graph: Option<RailwayGraph>,
    pub current_node: Option<NodeId>,
    pub target_node: Option<NodeId>,
    pub agent_rl: TrainAgentRL,
    pub trainer: Arc<RwLock<AgentTrainer<TrainAgentState>>>,
}

A reinforcement learning agent that controls a train in the simulation.

Fields

id: RailwayObjectId

The id of the railway object

railway_graph: Option<RailwayGraph>

The railway graph representing the train network.

current_node: Option<NodeId>

The current node

target_node: Option<NodeId>

The target node

agent_rl: TrainAgentRL

The reinforcement learning agent responsible for controlling the train.

trainer: Arc<RwLock<AgentTrainer<TrainAgentState>>>

The trainer responsible for training the reinforcement learning agent.

Implementations

impl TrainAgentAI

pub fn new(railway_graph: RailwayGraph, initial_state: TrainAgentState) -> Self

Creates a new TrainAgentAI with the given railway graph and initial state.

Arguments

• railway_graph - The railway graph representing the train network.
• initial_state - The initial state of the train agent in the simulation.

Returns

A new TrainAgentAI instance.

pub fn train(&mut self, iterations: usize)

Trains the reinforcement learning agent for the specified number of iterations.

Arguments

• iterations - The number of iterations to train the agent.

pub fn best_action(&self, state: &TrainAgentState) -> Option<RailMovableAction>

Returns the best action for the given state according to the trained reinforcement learning agent.

Arguments

• state - The current state of the train agent in the simulation.

Returns

The best action for the given state or None if no action can be selected.

pub fn observe( &mut self, current_node: NodeId, target_node: Option<NodeId>, speed_mm_s: Option<i32>, delta_distance_mm: Option<i32> )

Updates the current edge, target node (optionally), and calculates the new state by updating its distance using the shortest path distance while keeping the current speed constant.

Arguments

• current_edge - The current edge on which the train agent is.
• target_node - The optional target node the train agent is heading to.
• time_delta_ms - The time delta in milliseconds since the last update.

Trait Implementations

impl Clone for TrainAgentAI

fn clone(&self) -> TrainAgentAI

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for TrainAgentAI

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl DecisionAgent for TrainAgentAI

type A = RailMovableAction

The associated action type for this decision agent.

fn next_action(&self, _delta_time: Option<Duration>) -> Self::A

Returns the best action based on the current state of the simulation.

fn observe(&mut self, environment: &SimulationEnvironment)

Observes the current environment and updates the agent’s internal state.

fn as_any(&self) -> &dyn Any

Returns a reference to the Any trait for this object.

impl Default for TrainAgentAI

fn default() -> TrainAgentAI

Returns the “default value” for a type.