Class SimInf_model

The core class for storing the state, parameters, and results of a stochastic simulation in SimInf. This class holds the model definition (transition graphs, matrices), initial conditions, scheduled events, and the simulation output.

Slots

G

Dependency Graph (sparse matrix, class dgCMatrix). Indicates which transition rates need to be updated after a state transition occurs. A non-zero entry G[i, j] means that transition rate i must be recalculated if transition j occurs. This optimizes performance by avoiding unnecessary updates. Dimensions: \(N_t \times N_t\), where \(N_t\) is the number of transitions.

S

State Transition Matrix (sparse matrix, class dgCMatrix). Defines the change in the state vector for each transition. Executing transition j adds the column S[, j] to the state vector of the affected node. Dimensions: \(N_c \times N_t\), where \(N_c\) is the number of compartments.

U

Discrete State Result Matrix (integer matrix). Contains the number of individuals in each compartment for every node at each time point in tspan.

• U[, j]: State at time tspan[j].

• Rows are ordered by node, then by compartment:

  • Rows 1:Nc: Node 1.

  • Rows (Nc+1):(2*Nc): Node 2.

  • ... and so on.

Dimensions: \(N_n \times N_c \times \text{length(tspan)}\). Note: If the model was run with sparse output, this slot is empty.

U_sparse

Sparse Discrete State Result (sparse matrix, class dgCMatrix). Contains the simulation results if the model was configured for sparse output. The layout is identical to U, but stored as a sparse matrix to save memory. Note: Only one of U or U_sparse will contain data.

V

Continuous State Result Matrix (numeric matrix). Contains the values of continuous state variables (e.g., environmental pathogen load) for every node at each time point. Dimensions: \(N_n \times N_{ld} \times \text{length(tspan)}\), where \(N_{ld}\) is the number of local data variables (continuous states). Note: If sparse output was used, this slot is empty.

V_sparse

Sparse Continuous State Result (sparse matrix, class dgCMatrix). Contains the continuous state results if sparse output was enabled. Layout identical to V. Note: Only one of V or V_sparse will contain data.

ldata

Local Data Matrix (numeric matrix). Parameters specific to each node (e.g., node-specific transmission rates). Column ldata[, j] contains the local data vector for node j. Passed to transition rate functions and post-step functions. Dimensions: \(N_{ld} \times N_n\).

gdata

Global Data Vector (numeric vector). Parameters common to all nodes (e.g., global recovery rate). Passed to transition rate functions and post-step functions.

tspan

Time Span (numeric vector). Increasing time points where the state of each node is recorded.

u0

Initial State Matrix (integer matrix). Initial number of individuals in each compartment for every node. Dimensions: \(N_c \times N_n\).

v0

Initial Continuous State Matrix (numeric matrix). Initial values for continuous state variables for every node. Dimensions: \(N_{ld} \times N_n\).

events

Scheduled Events (SimInf_events). Object containing the schedule of discrete events (e.g., movements, births).

replicates

Number of Replicates (integer). Number of parallel replicates simulated for this model (used in filtering algorithms).

C_code

C Source Code (character vector). Optional C code defining the model's transition rates. If non-empty, this code is written to a temporary file, compiled, and loaded when run() is called. Typically generated by mparse.

See also

SimInf_model (constructor) for creating model objects. SIR, SEIR, SIS, SISe for compartment model constructors that automatically set up the required slots. mparse for creating custom models using a simple string syntax. run for executing the simulation. trajectory, prevalence, and plot,SimInf_model-method for extracting and visualizing results. SimInf_events for details on the event schedule structure.