graphenv.examples.tsp.tsp_state.TSPState

class TSPState(graph_generator, G=None, tour=[0])

Bases: graphenv.vertex.Vertex

Create a TSP vertex that defines the graph search problem.

Parameters

• generator – a function that creates a networkx graph

• G (Optional[networkx.classes.graph.Graph]) – A fully connected networkx graph.

• tour (List[int]) – A list of nodes in visitation order that led to this state. Defaults to [0] which begins the tour at node 0.

• graph_generator (Callable[[], networkx.classes.graph.Graph]) –

Return type

None

Methods

new	Convenience function for duplicating the existing node.
render	Optional method for rendering the current state of the environment.

Attributes

children	Gets the child vertices of this vertex.
info	Returns: An optional dictionary with additional information about the state
observation	Gets the observation of this vertex.
observation_space	Returns the graph env's observation space.
reward	Returns the graph env reward.
root	Returns the root node of the graph env.
terminal	Returns: True if this is a terminal vertex in the graph.

property info: Dict

Returns: An optional dictionary with additional information about the state

new(tour=[0], new_graph=False, **kwargs)

Convenience function for duplicating the existing node.

Parameters

• G – Networkx graph.

• tour (List[int]) – List of visited nodes.

Returns

New TSP state.

property observation_space: gymnasium.spaces.dict.Dict

Returns the graph env’s observation space.

Returns

Dict observation space.

render()

Optional method for rendering the current state of the environment.

Return type

Any

property reward: float

Returns the graph env reward.

Returns

Negative distance between last two nodes in the tour.

property root: graphenv.examples.tsp.tsp_state.TSPState

Returns the root node of the graph env.

Returns

Node with node 0 as the starting point of the tour, and generates a new graph using the given constructor