[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference rusure::math

Title:Mathematics at DEC
Moderator:RUSURE::EDP
Created:Mon Feb 03 1986
Last Modified:Fri Jun 06 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:2083
Total number of notes:14613

1663.0. "CMJ Problem 484" by BEING::EDP (Always mount a scratch monkey.) Fri Sep 11 1992 13:54

    Let A, B, and C be independent random variables uniformly distributed
    on [0,1].  What is the probability that all of the roots of the cubic
    equation x^3 + Ax^2 + Bx + C = 0 are real?
T.RTitleUserPersonal
Name
DateLines
1663.1Brute Force solution...SSAG::LARYLaughter & hope & a sock in the eyeMon Sep 14 1992 04:4432
1663.2an attempt at a start to solve thisSTAR::ABBASISpell checking is a family valueMon Sep 14 1992 05:3433
    all roots are real if

                   q^3+r^2 <= 0

    where 
            q= 1/3 B - 1/9 A^2
    and 
            r= 1/6 (BA-3C)- 1/27 A^3

         so ask, what is probability that  q^3+r^2 <=0   ?
    or after expanding and simplifying:
         what is probability that
    
         1/27 B^3 - 1/108 B^2 A^2 - 1/6 BCA + 1/4 C^2 + 1/27 CA^3 <=0  ?
    
    since B,A,C are all >= 0, then we must have
    
     1/27 B^3 + 1/4 C^2 + 1/27 CA^3 <= 1/108 B^2 A^2 + 1/6 BCA   --- (1)
     for all roots to be real.
    
    what is probability of (1) being true?
    rewrite (1) as
    
     4 B^3  + 27 C^2 + 4 CA^3 <= B^2 A^2 + 18 BCA                -- (2)
    so we ask, what is probability of (2) being true?
    
    I leave the rest as an excersise for the reader, I want to go and sleep ;-)
    
    /Nasser
    
    
    
1663.3SSAG::LARYLaughter &amp; hope &amp; a sock in the eyeMon Sep 28 1992 17:1015
1663.4VMSDEV::HALLYBFish have no concept of fire.Mon Sep 28 1992 19:012
    p.s., a hack BASIC Monte Carlo program, run for mega-trials,
    numerically verified Richie's answer.
1663.5Onward...SSAG::LARYLaughter &amp; hope &amp; a sock in the eyeMon Sep 28 1992 23:2320
Thanks for verifying it!

Now, of course, as is customary in MATH notes, we need to generalize the
problem. Let P(N) be the probability that a unary Nth degree polynomial
with uniform random coefficients in [0,1] has all real roots.

(1)	Prove that P(N) > 0 when N > 0. This looks like it would be easy if
	I knew enough algebra/analysis, but I can't get a handle on it...

(2)	Given (1),

		P(1) = 1
		P(2) = 1/12 (easy to derive)
		P(3) = 1/2880 (from note 1663.1)

		P(4) = ??

		P(N) = ?? (closed form function of N)

		Sum(N=1 to inf) P(N) = ????? (numeric value)