- The method starts with a point which can
choose at random or with some prejudice. We now wish to generate the
next point .
- Choose uniformly in a
neighborhood of , i.e., .
- Then examine and
- if , accept and
put and
- if , generate a new random number
**r**uniformly in .- if , set
- if , set

- is next random point (note it may be identical to ). Eventually, this produces a set of points distributed ``according to ''.

Geoffrey Fox, Northeast Parallel Architectures Center at Syracuse University, gcf@npac.syr.edu