Dynamic Mode Decomposition with control

Dynamic Mode Decomposition with Control is a method for generating an approximating linear differential equation in a chosen basis of observables. If X and Y are data matrices containing points of the same trajectory and U containing the exogenous inputs acting on that trajectory, then DMDc approximates

\[G = Y~\left[ \begin{array}{c} X \\ U \end{array} \right]^{\dagger} = \left[K ~B \right]\]

where $\dagger$ denotes the Moore-Penrose pseudoinverse and K is the approximation of the Koopman Operator and B the linear input map.

DMDc approximates discrete time systems with inputs $z$ of the form

\[u_{i+1} = K ~ u_{i} ~+ ~B ~ z_{i}\]

gDMDc approximates continuous time systems with inputs $z$ of the form

\[\frac{d}{dt}u = K_{G} ~ u + B ~ z\]

where $K_{G}$ is the generator of the Koopman Operator.

Functions

DataDrivenDiffEq.DMDc — Function

DMDc(X, U; B, alg)
DMDc(X, Y, U; B, alg)

Approximates a 'LinearKoopman' with the AbstractKoopmanAlgorithm 'alg' from the data matrices X and U or X, U and Y respectively. If only X is given, the data is split into X[:, 1:end-1] and X[:, 2:end]. B represents the matrix mapping the input onto Y, which may be known a priori.

Example

alg = DMDPINV()
koopman = DMDc(X, U, alg = alg)

koopman = DMD(X[:, 1:end-1], X[:, 2:end], U, alg = alg)

source

DataDrivenDiffEq.gDMDc — Function

gDMDc(X, Y, U; B, alg)
gDMDc(t, X, U; B, dt, alg, fdm, itp)

Approximates a 'LinearKoopman' with the AbstractKoopmanAlgorithm 'alg' from the data matrices X and U or X, U and Y, respectively. X should contain the state trajectory and Y the differential state trajectory.B represents the matrix mapping the input onto Y, which may be known a priori.

If no measurements of the differential state are available, gDMDc can be called with measurement time points t as the first argument. It will then create an interpolation using the interpolation method from DataInterpolations.jl defined in itp. The trajectory will then be resampled to equidistant measurements over time corresponding to the mean of diff(t) or dt, if given. The differential state measurements will be computed via 'FiniteDifferences.jl', given a FiniteDifferenceMethod in fdm.

Example

koopman = gDMDc(X, Y, U)

fdm = backward_fdm(5,1)
itp = CubicSpline
koopman = gDMD(t, X, U, fdm = fdm, itp = itp)

source