TRUNCATION IMPLICATIONS: PARTICLE SWARM OPTIMIZATION

Hi, I am trying to utilise PSO to optimise my problem which requires
integer variables (not binary)
I have defined a mapping but when I look at the literature e.g. on
task assignment, etc I find that most researchers utilise truncation
at the first iteration of the evolution of the population

Generate random position as follows:
Position(i,j) = [(UB(j) - LB(j))*rand + LB(j)]
then
Position(i,j) = round(Position(i,j))

where i indexes the population, j indexes the problem dimension
rand is just a uniform number between 0 and 1

Obviously written as 1 line in the code; but for clarity and to stress
my point, that people generally use truncation

My questions are: Does any one know of any other way to generate NON
BINARY integer variables?
Secondly: What if there are no upper bounds? How do we use PSO then?

Thanks

I have seen Maurice Clerc's work but find it very difficult to make
head or tails of it. I am sure its brilliant but I dont understand his
integer PSO method. Any advice?

Thank you all in advance

.