[MATH-1056] Small error in PoissonDistribution.nextPoisson() algorithm

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 3.2
• Fix Version/s: 3.3
• Assignee: Unassigned
• Reporter: Sean R. Owen
• Created: 03/Nov/13 14:19
• Updated: 19/May/14 15:13
• Resolved: 08/Nov/13 23:31

Description

Here's a tiny bug I noticed via static inspection, since it flagged the integer division. PoissonDistribution.java:325 says:

final double a1 = FastMath.sqrt(FastMath.PI * twolpd) * FastMath.exp(1 / 8 * lambda);

The "1 / 8 * lambda" is evidently incorrect, since this will always evaluate to 0. I rechecked the original algorithm (http://luc.devroye.org/devroye-poisson.pdf) and it should instead be:

final double a1 = FastMath.sqrt(FastMath.PI * twolpd) * FastMath.exp(1 / (8 * lambda));

(lambda is a double so there is no int division issue.) This matches a later expression.

I'm not sure how to evaluate the effect of the bug. Better to be correct of course; it may never have made much practical difference.