Polygon in optimized constraint representation (HPolygonOpt)

LazySets.HPolygonOptType
HPolygonOpt{N, VN<:AbstractVector{N}} <: AbstractHPolygon{N}

Type that represents a convex polygon in constraint representation whose edges are sorted in counter-clockwise fashion with respect to their normal directions. This is a refined version of HPolygon.

Fields

  • constraints – list of linear constraints, sorted by the normal direction in counter-clockwise fashion
  • ind – index in the list of constraints to begin the search to evaluate the support function
  • sort_constraints – (optional, default: true) flag for sorting the constraints (sortedness is a running assumption of this type)
  • check_boundedness – (optional, default: false) flag for checking if the constraints make the polygon bounded; (boundedness is a running assumption of this type)
  • prune – (optional, default: true) flag for removing redundant constraints

Notes

This structure is optimized to evaluate the support function/vector with a large sequence of directions that are close to each other. The strategy is to have an index that can be used to warm-start the search for optimal values in the support vector computation.

The option sort_constraints can be used to deactivate automatic sorting of constraints in counter-clockwise fashion, which is an invariant of this type. Alternatively, one can construct an HPolygonOpt with empty constraints list, which can then be filled iteratively using addconstraint!.

Similarly, the option prune can be used to deactivate automatic pruning of redundant constraints.

Another type assumption is that the polygon is bounded. The option check_boundedness can be used to assert this. This option is deactivated by default because we explicitly want to allow the iterative addition of the constraints, and hence one has to initially construct an empty list of constraints (which represents an unbounded set). The user has to make sure that the HPolygonOpt is not used before the constraints actually describe a bounded set. The function isbounded can be used to manually assert boundedness.

source
LazySets.σMethod
σ(d::AbstractVector, P::HPolygonOpt;
  [linear_search]::Bool=(length(P.constraints) < BINARY_SEARCH_THRESHOLD))

Return the support vector of an optimized polygon in a given direction.

Input

  • d – direction
  • P – optimized polygon in constraint representation
  • linear_search – (optional, default: see below) flag for controlling whether to perform a linear search or a binary search

Output

The support vector in the given direction. The result is always one of the vertices; in particular, if the direction has norm zero, any vertex is returned.

Algorithm

Comparison of directions is performed using polar angles; see the overload of <= for two-dimensional vectors.

For polygons with BINARY_SEARCH_THRESHOLD = 10 or more constraints we use a binary search by default.

source
LazySets.translateMethod
translate(P::HPolygonOpt, v::AbstractVector; share::Bool=false)

Translate (i.e., shift) an optimized polygon in constraint representation by a given vector.

Input

  • P – optimized polygon in constraint representation
  • v – translation vector
  • share – (optional, default: false) flag for sharing unmodified parts of the original set representation

Output

A translated optimized polygon in constraint representation.

Notes

The normal vectors of the constraints (vector a in a⋅x ≤ b) are shared with the original constraints if share == true.

Algorithm

We translate every constraint.

source

Inherited from ConvexSet:

Inherited from AbstractPolytope:

Inherited from AbstractPolygon:

Inherited from AbstractHPolygon: