[Ns-bugs] [Bug 181] New: Normal Variable Infinite Value & Bounds

[bugzilla-daemon@nsnam-www.ece.gatech.edu]

http://www.nsnam.org/bugzilla/show_bug.cgi?id=181

           Summary: Normal Variable Infinite Value & Bounds
           Product: ns-3
           Version: unspecified
          Platform: All
        OS/Version: All
            Status: NEW
          Severity: minor
          Priority: P3
         Component: simulation core
        AssignedTo: ns-bugs at isi.edu
        ReportedBy: tjkopena at cs.drexel.edu


Two quick points:

- NormalVariable::INFINITE_VALUE is no longer defined.  NormalVariableImpl
defines and uses a value, but NormalVariable declares and uses a static const
INFINITE_VALUE to define some default values, without ever defining it.  It
should be defined, or probably more correctly, the two methods that use it
should be changed such that there is no default but instead one version w/ no
third parameter, and another w/ the bounds parameter, so that the underlying
implementation can work with defaults however it wants.

- I am not totally sure what this would do to the current algorithm, but it
seems reasonable that the valid region should be mean +/- bound rather than
-bound to bound.  This makes it more convenient, for example, to create a
normal variable w/ distribution 2, 0.1 bounded by 1--3, i.e. to jitter some
periodic packet generation.  Taking a glance, it seems like this should not
affect the algorithm, but I am not an RNG expert.

Thx


-- 
Configure bugmail: http://www.nsnam.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.