| Arun
Iyer |
Prove that tan2 (1o )+tan2 (2o )+..+tan2 (89o ) is an integer love arun P.S -> Answers may be written in white |
||
| Sarah
Sarah |
Arun I tried it out on Excel and got something.3333, not an integer. |
||
| Arun
Iyer |
sorry sarah! my mistake!( you could prolly sue me for taxing your brain for nothing! but please don't )the question restated, Prove that tan2 (1o )+tan2 (3o )+..+tan2 (89o ) is an integer love arun |
||
| Arun
Iyer |
hmm, if anyone is trying then here is a vague hint, "De Moivre ". love arun |
||
| Graeme
McRae |
Here's what I've discovered so far... I have discovered that if you divide the interval [0,90º] into n equal intervals, and sum the squares of tan of the angle at the center of each of the intervals, then this sum is 2n^2-n. Arun's problem is a case in point, where n=45. The sum of squares of tans is 2n^2-n, which is 4005 when n=45. Other similar problems would be to prove the sum of tan^2 15º + tan^2 45º + tan^2 75º = 15, or that tan^2 9º + tan^2 27º + tan^2 45º + tan^2 63º + tan^2 81º = 45. --Graeme |
||
| Arun
Iyer |
wow that's a pretty neat result Graeme!! Is it easy to prove? love arun |
||
| Graeme
McRae |
Arun, I think it should be fairly easy to prove... The angles in the range between 0º and 90º are just the solutions in the first quadrant to the equation cos 180x + 1 = 0 Expanding cos 180x in terms of cos x, we get cos 180x = 1 - 16200cos^2 x + ... where the "..." represent a whole bunch of other even powers of cos x, with humongous coefficients, the values of which, luckily, don't matter to us! Plugging this into the first equation above, we get the polynomial equation, 0 = 2 - 16200cos^2 x + ... I'm sure you'll recognize 16200 as half the square of 180 -- this is important for the general case. Since sec is the reciprocal of cos, this polynomial equation can be expressed as 0 = 2sec^180 x - 16200sec^178 x + ... Let u = sec^2 x, so now we have 0 = 2u^90 - 16200u^89 + ... The sum of the 90 roots of this polynomial is 8100, but we're only looking for the sum of the 45 values of u representing angles in the first quadrant, so we divide that in half, giving us 4050. To complete the proof, use 4n in place of 180 to find the sum of squares of secants is 2n2 , and then subtract n to get the sum of squares of tangents. --Graeme |
||
| Arun
Iyer |
Graeme, cos(180*2)+1 = 0 ? love arun |
||
| Graeme
McRae |
Arun, No, I meant the angles of your problem are the solutions in the first quadrant to the equation cos 180x + 1 = 0 Solutions to this equation are 1º, 3º, etc. --Graeme |
||
| Arun
Iyer |
yes graeme, i just realised that as well. Very Nicely done!!! Also realised that your proof is essentially the same as mine except that i start with ::De Moivre :: and you have started with a nice simple equation. just to add to flavours i would give the proof to Graeme's generalisation (though it is essentially same as Graeme's proof , just a tad bit different) :: expand RHS and equate real parts, The 's in question are roots of the equation, hence, divide by on both sides and let , hence sum of roots :: love arun |
||
| Andre
Rzym |
It ought to be possible to do something similar using complex numbers. If then So construct a poly that satisfies (we can have the angles go through the full 360 degrees rather than 90 degrees - this just quadruples the sum you require). If you substitute the resulting poly in should allow you to evaluate the equation for above, summed over all required angles. Hopefully this makes some sense - if not I'll be forced to work through the gory details! Andre |
||
| Kerwin
Hui |
Andre, it should be . Kerwin. |
||
| Arun
Iyer |
Sorry Andre, but i did not get this statement, Quote: "..So construct a poly that z satisfies.." I did not get the objective behind this!? love arun |
||
| Andre
Rzym |
Arun, sorry for the delay - I'll post something more complete on Sunday. Andre |
||
| Andre
Rzym |
Arun, Consider (for simplicity) the case when there are just two terms. We want where Now if then Furthermore, if (A) then , , ..., So actually, where the sum is over the roots of the poly above. If we write then (B) Now each root of (B) corresponds to 2 roots of (A) (a root and the same but the opposite sign). Therefore Where the sum is over the roots of(B), of which there are 4: It is at this point that the construction of the polynomial becomes significant, becausethe coefficients tell us about the sums of the roots taken 1, 2, etc. at a time. In particular, the coefficient of in (B) is , therefore and therefore Finally, (sum of roots taken1 at a time) 2(sum of roots taken2 at a time) Hence Having points in the quadrant rather than just 2 should change the computations very little. Andre |
||
| Andre
Rzym |
So to complete the proof when there are terms in the sum of squares, with defined as above, we would have where the sum is over the roots of the polynomial Therefore where the sum is over the roots of Therefore Andre |
||
| Arun
Iyer |
(Sorry to post so late .. have my exams going on!) Btw, nice proof! The translation from z to u is the neatest trick.(Not sure i would have thought of that!!) love arun |
||
| Andre
Rzym |
Arun, There is an interesting consequence of your summation problem: We are all agreed that ![]() From which ![]() Now tan a > a [for 0 < a < pi/2] so cot a < 1/a. Therefore, ![]() Multiply both ends by pi2 /16n2 : ![]() Now take the limit as n -> infinity ![]() Now it's possible to create a similar inequality from trigonometric identities, but the other way round. In other words we are forced to conclude that: ![]() which is a nice relationship. The usual way to prove this sort of thing is via Fourier Transforms, but the approach here is an interesting alternative. Andre |
||
| Arun
Iyer |
Beautiful!!(well spotted Andre!!) This is too cool a result. Ofcourse from here we could go ahead and prove zeta(2)(which is simply 4/3 of the above result ,putting the convergence arguments under the rug for the moment). I know this is a mere coincidence but still i wonder whether higher powers are possible?? love arun |
||
| Andre
Rzym |
Higher powers ought to be possible. For example, I think ![]() from which we ought to be able to deduce an inequality analogous to the one above. I haven't experimented, but would guess that this is not going to work for odd powers, but will work for any even powers. Andre |
||
| Arun
Iyer |
Andre, your expression to fourth power is correct and it is surprisingly easy to prove..(barring a few clumsy algebra). And you know what, you may be even right about odd powers and even powers!!! love arun |