Randomizing

NAME

: randomize, uniform, geometric1, geometric2, normal, randfunc - Randomizing routines

SYNTAX

:

      int randomize ( int x[3] )

      int uniform ( int l, int u )

      int geometric1 ( int l, int u, int m )

      int geometric2 ( int l, int u, int m )

      normal ( int m, int u, int v )

      float randfunc ()

DESCRIPTION

:
These routines use the system function random in order to produce a random number according to a specific probability distribution function.

• randomize returns a step or delay sample according to the time distribution function used at the specific execution and recorded to the global variable TIMEDISTR, and also the argument array x which contains the parameters of the distribution function. So in order to obtain a step sample for process i you can write: $step\_sample = randomize ( PCB[ i ].step\_dur)$ and for a delay sample for process $i$ and port j: $delay\_sample = randomize ( PCB[ i ].adjust[ j ].delays)$
• uniform returns a random integer sample between l and u with equal probability.
• geometric1 returns a random integer sample between l and u and according to a geometric distribution with mean value m.
• geometric2 returns a random integer sample between l and u and according to a geometric distribution with mean value m. The difference from geometric1 is that in geometric1 there is a cutoff of the values greater than u (for these cases geometric1 returns u) while geometric2 retries for finite number of tries and finally chooses a samle between l and u like as uniform does.
• normal returns a random integer sample between 1 and u according to a Gaussian distribution with mean value m and variance v.
• randfunc returns a random real sample between 0 and 1 according to a uniform distribution.

SEE ALSO: Flexible Timing Assumptions (Section 4.3.1), System structure (Section 5.2) and Distr.h.