BFFPSV18

ReachabilityAnalysis.BFFPSV18Type
BFFPSV18{N, ST, AM, IDX, BLK, RBLK, CBLK} <: AbstractContinuousPost

Implementation of the reachability method for linear systems using block decompositions.

Fields

  • δ – step-size of the discretization
  • approx_model – (optional, default: Forward) approximation model; see Notes below for possible options
  • vars – vector with the variables of interest
  • block_indices – vector of integers to index each block that contains a variable of interest
  • row_blocks – vector of integer vectors to index variables associated to blocks of interest
  • column_blocks – vector of integer vectors to index variables in the partition
  • lazy_initial_set – (optional, default: false) if true, use a lazy decomposition of the initial states after discretization
  • lazy_input – (optional, default: false) if true, use a lazy decomposition of the input set after discretization
  • sparse – (optional, default: false) if true, assume that the state transition matrix is sparse
  • view – (optional, default: false) if true, use implementation that uses arrays views

matrix is sparse

See the Examples section below for some concrete examples of these options.

Notes

This algorithm solves the set-based recurrence equation $X_{k+1} = ΦX_k ⊕ V_k$ by using block decompositions. The algorithm was introduced in [BFFPSV18].

Comments about some fields:

  • N – number type of the step-size, e.g. Float64
  • ST – set representation used; this is either a concrete LazySet subtype, eg. Interval{Float64}, or a tuple of concrete LazySet subtypes that is commensurate with the partition

The default approximation model is:

Forward(sih=:concrete, exp=:base, setops=:lazy)

TODO:

  • clarify assumption about contiguous blocks

Examples

References

This algorithm is essentially an extension of the method in [BFFPSV18]. Blocks can have different dimensions and the set representation can be different for each block.

For a general introduction we refer to the dissertation [SCHI18].

Regarding the approximation model, by default we use an adaptation of the method presented in [FRE11].

source