Simulation Output

If you run a simulation

data = simulate(my_model, T = T)

the object data will contain all relevant simulation output as a NamedTuple.

Fields

The output object mimics the structure of the respective model and has the following fields.

Model information:

• model_id the integer ID assigned to the model
• Δt the sampling time of the model, or nothing for pure continuous-time models

Continuous-time output:

• tcs vector of time values at which the continuous-time state was recorded
• xcs state trajectory produced by fc. nothing if the model has no continuous-time dynamics.
• ycs output trajectory produced by gc. nothing if the model has no continuous-time output function.

Discrete-time output:

• tds vector of time values at which the discrete-time state was recorded
• xds state trajectory produced by fd. nothing if the model has no discrete-time dynamics.
• yds output trajectory produced by gd. nothing if the model has no discrete-time output function.
• wds random draw trajectory produced by wd. nothing if the model has no random draw function.

Submodels:

• models contains the outputs of all submodels, mirroring the structure of the models field in the original model definition.

Time-Series Format

If a state or output is an AbstractVector of numbers (the common case), SimpleSim.jl stacks all time steps into a matrix after the simulation finishes. Each row corresponds to one time step and each column corresponds to one state or output dimension.

data.tcs          # Vector of length N
data.xcs          # Matrix of size N × nx
data.ycs          # Matrix of size N × ny

If the state or output is a scalar or a non-numeric type it is returned as a plain vector with one entry per time step.

Accessing Submodel Output

Output of submodels can be accessed via the models field. For example, if the top-level model has a continuous-time submodel called my_submodel, then its output can be accessed via

data.models.my_submodel.ycs