ExpmV.expmv
— Methodexpmv(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
orStepRangeLen
objectA
: an × n
real or complex sparse matrixb
: ann
-vectorM = []
: manually set the degree of the Taylor expansionprecision = "double"
: can be"double"
,"single"
or"half"
.shift = false
: set totrue
to apply a shift in order to reduce the norm of A (see Sec. 3.1 of the paper)full_term = false
: set totrue
to evaluate the full Taylor expansion instead of truncating when reaching the required precision