Inconsistent behavior of GLGM06 across (in)homogeneous cases

[schillic]

In the homogeneous case, we use an algorithm with wrapping effect:

Reachability.jl/src/ReachSets/ContinuousPost/GLGM06/reach.jl

Line 19 in 29f5a2c

HR_next = linear_map(Φ, HR[k-1].X)

In the inhomogeneous case, we use an algorithm without wrapping effect (except in the inputs):

Reachability.jl/src/ReachSets/ContinuousPost/GLGM06/reach.jl

Line 51 in 29f5a2c

Xk = minkowski_sum(linear_map(Φ_power_k, Ω0), Wk₊)

[mforets]

In the homogeneous case, we use an algorithm with wrapping effect:

in this scenario, the linear map is exact, so i don't expect to see a difference between recursive and not. in fact the call to reduce_order in the following line is not needed.

[mforets]

do reopen if you disagree :)