The Wigner function (2.2) is defined with the Wigner transform (2.1) calculated over an infinite range. However, the finite dimensions of the simulation domain impose bounds on the variables defined by (2.3) such that the maximum value that s can attain is limited by the dimensions of the device (Ldev). Therefore, the Wigner function (2.2) must be calculated over finite dimensions:
A continuous function f(s) can be written as a Fourier series:
where the Fourier coefficients are given by Therefore, a finite value of L requires k to become discretized to form a complete orthogonal basis set , where q is an integer multi-index and Δk = , which denotes the resolution of the discretized wavevector.From the above considerations, the discrete Fourier transform (DFT) follows, such that
The Wigner potential (which may also be time-dependent) is defined accordingly as
The equations (2.35) - (2.36) define the semi-discrete Wigner equation, which has been the subject of detailed mathematical analyses [66, 67].