[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

282.0. "Triangles w/ int. sides + area" by RAINBO::GRANT () Thu May 09 1985 23:26

All these puzzles and problems are about triangles with integral sides and 
integral areas, called "integral triangles" for short.

1. The areas of integral triangles are always even: why?
 
2. As a related problem, or as a warmup, try showing that all pythagorean 
triangles are integral triangles.

3. There are only 5 integral triangles whose sides add up to their area.
Find these, and as a bonus, show they are the only ones.

4. The larger integral triangles sometimes have areas that are integral 
multiples of the sums of their sides.  For any integer M there are a finite 
number K>1 of integer triangles whose area = perimeter * M.  I know of no 
way, short of enumeration, of evaluating f(M) = K.  Maybe someone can find one.
The first few terms:

      M      K
      1      5
      2      15
      3      33

I haven't proved yet that I have f(2) and f(3) right: I only have upper and 
lower bounds.  There may be triangles that satisfy the criteria where all 
the sides are > 100. (Although I doubt it) 

5. Sometimes different integral triangles have the same area.  What's the 
smallest area this is true for, and what are the two integral triangles that 
have this area?

6. As in the solution to #5, sometimes these equal area triangles have two 
sides out of three in common. Let's call these "linked equiareal triangles."
Even excluding multiples of smaller l.e.t's, there are an infinite number of 
these.  (I'm pretty sure, anyway: can anyone prove it?)  

7. In l.e.t,'s, one of the two lengths that doesn't match up is always the 
largest length of its triangle.  How come?

8. The most acute integral triangle I have found so far has sides of 3, 160, 
and 163.  I imagine that there are always integral triangles that are more 
acute, but I haven't investigated much. Are there? Is there a limit?

9. The most obtuse integral triangle is probably the same as the most acute 
one, right?

Some of these I have answers to, some not. As always, contributions are 
welcome from all.

-Jim Grant, Littleton MA.
       
T.RTitleUserPersonal
Name
DateLines
282.1RAINBO::GRANTFri May 10 1985 01:4710
Editorial corrections:

8. Sides of 160,163, and 3 do not make a triangle in my kind of universe.
I meant to say 160,162, and 3. Even that misses being an integral triangle by 
a few thousandths.  Anyway, in the meantime, I found an integral triangle 
with sides 3,865, and 866.  Very acute.  An area of 1224.

9. This question should have the semi-universal bulletin board smile figure
appended to it, like so:     :-)

282.2ALIEN::POSTPISCHILFri May 10 1985 16:1674
1)	Given:	Sides a, b, and c of a triangle, and its area, A, are all
		integers.

	Prove:	A is even.

	One formula for the area of a triangle is:

		A = sqrt( s * (s-a) * (s-b) * (s-c) ),

	where s = (a+b+c)/2.  Call the product inside the square root function
	P.  Since P = A*A, P is an integer.

	Lemma:	s is an integer.

		Suppose s be not an integer (this is proper English, see note
		68 in SUMMIT""::SYS$NOTES:JOYOFLEX).  Then s is in the form
		i0/2, where i0 is an odd integer (i0 = a+b+c).  Clearly, s-a,
		s-b, and s-c also have forms i1/2, i2/2, and i3/2, for some
		odd integers i1, i2, and i3.  So P = i0*i1*i2*i3/16.  But
		i0*i1*i2*i3 must be odd, so P is not an integer, which is a
		contradiction.  Therefore s is an integer.
	
	Lemma:	P is even.

		Suppose P be odd.  Then none of s, s-a, s-b, and s-c could
		be even integers.  Since they are all integers, they must all
		be odd.  If s is odd and s-a is odd, then a must be even.
		Similarly, b and c must be even.

		If they are all even, then a = 2i, b = 2j, and c = 2k, for some
		integers i, j, and k.  Expanding P reveals

			P = 2(i*i*j*j + j*j*k*k + k*k*i*i) - i**4 - j**4 - k**4.

		Consider the remainder when P is divided by 4 when i, j, and k
		take on even or odd values.  Because i, j, and k are symmetric
		in the above formula, it is only necessary to consider one case
		where they are all even, one case where one is odd, one case
		where two are odd, and so on.

			i j k | formula, shown as congruences mod 4 | P mod 4
			0 0 0 | 2(0+0+0)-0-0-0                      | 0
			0 0 1 | 2(0+0+0)-0-0-1                      | 3
			0 1 1 | 2(0+1+0)-0-1-1                      | 0
			1 1 1 | 2(1+1+1)-1-1-1                      | 3

		So P is congruent to either 0 or 3 modulo 4.  If it is
		congruent to 0, it is even, which contradicts the supposition
		that P is odd.  If it is congruent to 3, then A is not an
		integer, because there is no integer whose square is congruent
		to 3 modulo 4 (integers congruent to 0, 1, 2, and 3 have
		squares congruent to 0, 1, 0, and 1, modulo 4).  This is also
		a contradiction.

		Therefore the supposition that P is odd is false.

	Since P is an even integer and A is an integer, then A is also even
	(otherwise squaring A, an odd integer, would yield P, an even integer). 


2)	Is a Pythagorean triangle a right triangle with integral sides?  If
	so, one of the legs is necessarily even, so the area is obviously an
	integer.

	To see that one side is even, suppose they be both odd.  Let the length
	of the hypoteneuse be denoted by a and the lengths of the legs by b
	and c.  Then a*a = b*b + c*c.  Since b and c are odd, b*b and c*c are
	odd, and b*b + c*c is divisible by two but not by four.  Clearly,
	no integer has a square divisible by two but not by four.  But a is
	known to be an integer, so this is a contradiction.  Therefore b and
	c are not both odd.


				-- edp (WHOAREYOU note 329)
282.3TOOLS::STANSat May 25 1985 21:1126
Of course you all know the formula for generating the sides of
all Pythagorean triangles:

		k(2mn)
		k(m^2-n^2)
		k(m^2+n^2)

where gcd(m,n)=1, m>n.

For quite a while, I have been researching the question of whether
there is a formula that generates all Integral Triangles
(also known as Heronian Triangles in the literature).  Dickson
hedges the question in his monumental work, the History of the Theory
of Numbers.  I have found several purported formulas in the literature,
some of which are wrong.  I believe the following is a valid formula:

		kmn(p^2+q^2)
		kpq(m^2+n^2)
		k(mq+np)(mp-nq)

where gcd(m,n)=1 and gcd(p,q)=1, mp>nq,

although the published proof seemed wrong.
Note that letting p=q=1 yields the formula for Pythagorean triples.

I hope you find this formula useful.  Let me know if it misses any solutions.
282.4R2ME2::STANSun May 26 1985 18:411
See also note 70 for a, yet unsolved, problem about Heronian triangles.
282.5RAINBO::GRANTTue May 28 1985 21:538
The formula in .3 seems to miss some triangles.  (5,5,6) is Heronian, with an 
area of 12, and cannot be generated by the formula if k is required to be an 
integer.  It does come from the formula when m=p=2,  n=q=1,  and k = 1/2.

If this is a valid use of the formula, then what is the domain of k?  k=1/2 
does not always produce integral solutions.

-Jim
282.6TOOLS::STANTue May 28 1985 23:021
Oops, another candidate down the drain.
282.71-9JOBURG::BUCHANANMon Oct 30 1995 17:22181
282.8I'll mail Stan with the formulaJOBURG::BUCHANANThu Nov 02 1995 07:25109