expmv(t, A, b; <keyword arguments>)

Returns the matrix-vector product $exp(t A) b$ where A is a $n × n$ sparse real or complex matrix, b is a vector of $n$ real or complex elements and t is a parameter (or a StepRangeLen object representing a range of values).

Arguments

• t: Number or StepRangeLen object
• A: a n × n real or complex sparse matrix
• b: an n-vector
• M = []: manually set the degree of the Taylor expansion
• precision = "double": can be "double", "single" or "half".
• shift = false: set to true to apply a shift in order to reduce the norm of A (see Sec. 3.1 of the paper)
• full_term = false: set to true to evaluate the full Taylor expansion instead of truncating when reaching the required precision