Trait TreeNode 

pub trait TreeNode {
    type State;
    type Action;
    type Cost;

    // Required methods
    fn state(&self) -> &Self::State;
    fn parent(&self) -> Option<(usize, &Self::Action)>;
    fn cost(&self) -> Self::Cost;
    fn queue_evaluation(&self) -> Self::Cost;
    fn queue_bias(&self) -> Option<Self::Cost>;
}

Required Associated Types

type State

The type used to describe the state of the node.

type Action

The type of action that can be taken from one node to another.

type Cost

The type used to decide whether to prefer one node over another. Lower values are preferable.

Required Methods

fn state(&self) -> &Self::State

Get the state of the node.

fn parent(&self) -> Option<(usize, &Self::Action)>

If the node has a parent, get the identity of that parent and the action used to arrive from it.

fn cost(&self) -> Self::Cost

Get the actual cost of arriving at this node from its initial state.

fn queue_evaluation(&self) -> Self::Cost

Evaluate this node for its placement in the search queue. For an uninformed node this would simply return its aggregated cost. For an informed node this would return its aggregated cost plus its remaining cost estimate.

fn queue_bias(&self) -> Option<Self::Cost>

Give a bias to this node. When queue_evaluation is exactly equal, the node will be ordered by this bias instead. Higher bias will push it later in the queue. For an informed node this could be its remaining cost estimate.

Implementors

impl<State, Action, Cost> TreeNode for mapf::algorithm::a_star::Node<State, Action, Cost>

where Cost: Clone + Add<Cost, Output = Cost>,

type State = State

type Action = Action

type Cost = Cost

impl<State, Action, Cost: Clone> TreeNode for mapf::algorithm::dijkstra::forward::Node<State, Action, Cost>

type State = State

type Action = Action

type Cost = Cost