Beruflich Dokumente
Kultur Dokumente
AND
EXPECTATIONS
RANDOM
INTERFERENCE
TYPE
L.L . Campbell1.
OF
VARIABLES
OF
where
{X1,X2,.. . }
is
sequence
of
independent
In
the
applications
envisaged
converge absolutelx.
If they satisfy the stronger
condition p = O(r ) for r<M then the distribution
k
here,
Acknowledgement
This research was supported by the Natural
Sciences and Engineering Research Council of Canada
through Grants A2151 and A3391, and through a
scholarship.
error
probability
calculations
on
one typically wishes to
calculate E[g(Z)] for some smooth function g . Because
of the special nature of the random variable, we
cannot employ either of the two standard tools,
involving an integral of a density function, o r an
infinite
series with
weights
equal
to
the
It is possible to
probabilities, to compute E[g(Z)I.
express Z as a discontinuous function defined on the
unit interval [0,1) and to write E[g(Z)I as an
integral on this interval.
This integral can be
approximated arbitrarily well by a trapezoidal
integration rule. The only condition which must be
assumed on the coefficients is that the series Zp
communication
channels,
References
310