nwmodel class

Captures a model of a dynamic network, in which parameters governing the dynamic relationships between the elements of the network can be estimated. It defines the mapping of all internal nodes, excitation signals and noise sources to the internal nodes of the network (as denoted in $w(t) = G(q)w(t) + R(q)r(t) + H(q)e(t)$ for directed networks). It is used in full network identification and simulation. The nwmodel class is a special case of the network predictor model (nwpredmodel) class, in which all excitations of the network are used as input signals, and all nodes are used as input and output signals of the predictor model. All properties and methods of nwpredmodel are also present in instances of this class.

Construction

• model = nwmodel(NetworkStructure)  creates a full network identification model based on the topology in the provided LabelledAdjStruct argument.
• model = nwmodel(G,T,H)  creates a parameterized full network identification model based on the G, T and H transfer functions. The underlying network structure is inferred from these objects.
• model = nwmodel(__,Name=Value)  creates a full network identification model with additional properties set according to the provided values. The same properties can be set through this method as in nwpredmodel.

Properties

See nwpredmodel for a list of properties shared with the nwpredmodel class.

• L

  positive scalar

  Number of nodes.

• K

  positive scalar

  Number of excitations.

• p

  positive scalar

  Number of noise signals.

Methods

See nwpredmodel for a list of methods shared with the nwpredmodel class.

• simulate

  Simulate identified network model.

  Syntax

  • w = simulate(model,r,e,t)  simulates the full network and returns the node signals w in response to the excitations r and noise signals e over the time samples t.
  • w = simulate(model,data,e)  simulates the full network and returns the node signals w in response to the excitation signals in data and noise signals e. The excitation signals in the data are matched to the labels in the network model, and the time samples are derived from the data object. Note that node signals in the data object are ignored.

  Input arguments

  • model

    nwmodel object

  • rdata

    $[N,n_{R}]$ double array	nwdata object

    Input excitation signals/data object.

  • e

    $[N,n_{E}]$ double array

    Input noise signals.

  • t

    $[N,1]$ double array

    Time samples.

  Output arguments

  • w

    $[N,n_{Y}]$ double array

    Simulated node signals