HeterogeneousNetwork

experimental.network_dynamics.HeterogeneousNetwork(
    graph,
    groups,
    routes=None,
    history=None,
)

Compose different dynamics groups on one shared square graph.

Args: graph: Shared dense or sparse graph. Its node order is canonical for all group assignments and signal routes. groups: Mapping from unique names to NodeGroup instances. Groups must form an exhaustive, non-overlapping partition of graph nodes. routes: Optional mapping from unique names to SignalRoute instances. history: Optional HeterogeneousSolution used to warm-start every group’s integrated state and delayed route signal history.

Attributes: graph: Shared graph instance. groups: Normalized group dictionary. routes: Normalized route dictionary. group_names: Canonically sorted group names. route_names: Canonically sorted route names. group_nodes: Normalized node-index tuples by group. n_nodes: Number of nodes in the shared graph.

Notes: Group membership and sparse topology are static after prepare(). Group/route parameters, graph values, delays within prepared capacity, noise, external inputs, initial states, and route histories remain live.

Methods

Name	Description
initial_state_for	Return the group-local initial state, including an active warm start.
update_history	Use a heterogeneous result to warm-start state and route history.

initial_state_for

experimental.network_dynamics.HeterogeneousNetwork.initial_state_for(name)

Return the group-local initial state, including an active warm start.

update_history

experimental.network_dynamics.HeterogeneousNetwork.update_history(solution)

Use a heterogeneous result to warm-start state and route history.

All integrated state variables must be present. This mirrors Network.update_history: variables-of-interest may be reordered, but auxiliary-only or reduced results cannot initialize a continuation.

# HeterogeneousNetwork { #tvboptim.experimental.network_dynamics.HeterogeneousNetwork }

```python
experimental.network_dynamics.HeterogeneousNetwork(
    graph,
    groups,
    routes=None,
    history=None,
)
```

Compose different dynamics groups on one shared square graph.

Args:
    graph: Shared dense or sparse graph. Its node order is canonical for all
        group assignments and signal routes.
    groups: Mapping from unique names to ``NodeGroup`` instances. Groups
        must form an exhaustive, non-overlapping partition of graph nodes.
    routes: Optional mapping from unique names to ``SignalRoute`` instances.
    history: Optional ``HeterogeneousSolution`` used to warm-start every
        group's integrated state and delayed route signal history.

Attributes:
    graph: Shared graph instance.
    groups: Normalized group dictionary.
    routes: Normalized route dictionary.
    group_names: Canonically sorted group names.
    route_names: Canonically sorted route names.
    group_nodes: Normalized node-index tuples by group.
    n_nodes: Number of nodes in the shared graph.

Notes:
    Group membership and sparse topology are static after ``prepare()``.
    Group/route parameters, graph values, delays within prepared capacity,
    noise, external inputs, initial states, and route histories remain live.

## Methods

| Name | Description |
| --- | --- |
| [initial_state_for](#tvboptim.experimental.network_dynamics.HeterogeneousNetwork.initial_state_for) | Return the group-local initial state, including an active warm start. |
| [update_history](#tvboptim.experimental.network_dynamics.HeterogeneousNetwork.update_history) | Use a heterogeneous result to warm-start state and route history. |

### initial_state_for { #tvboptim.experimental.network_dynamics.HeterogeneousNetwork.initial_state_for }

```python
experimental.network_dynamics.HeterogeneousNetwork.initial_state_for(name)
```

Return the group-local initial state, including an active warm start.

### update_history { #tvboptim.experimental.network_dynamics.HeterogeneousNetwork.update_history }

```python
experimental.network_dynamics.HeterogeneousNetwork.update_history(solution)
```

Use a heterogeneous result to warm-start state and route history.

All integrated state variables must be present. This mirrors
``Network.update_history``: variables-of-interest may be reordered, but
auxiliary-only or reduced results cannot initialize a continuation.