07-401/Andrei Litvin Poly Solution

Here is what I have so far in terms of polynomials:

Polynomial with root ${\sqrt {2}}$:

${x}^{2}-2$ (duh)

Polynomial with root ${\sqrt[{5}]{7}}$:

$x^{5}-7$ (duh as well)

Polynomial with root ${\sqrt {2}}+{\sqrt[{5}]{7}}$:

$x^{10}-10*x^{8}+40*x^{6}-14*x^{5}-80*x^{4}-280*x^{3}+80*x^{2}-280*x+17$ (more interesting!)

Polynomial with root ${\sqrt[{3}]{{\sqrt {2}}+{\sqrt[{5}]{7}}}}$:

$x^{30}-10*x^{24}+40*x^{18}-14*x^{15}-80*x^{12}-280*x^{9}+80*x^{6}-280*x^{3}+17$ (similar to the above)

Polynomial with root ${1} \over {\sqrt[{3}]{{\sqrt {2}}+{\sqrt[{5}]{7}}}}$:

$17*x^{30}-280*x^{27}+80*x^{24}-280*x^{21}-80*x^{18}-14*x^{15}+40*x^{12}-10*x^{6}+1$ (just invert the above .. quite the same as the above 2)

Polynomial with root ${\sqrt {5}} \over {\sqrt[{3}]{{\sqrt {2}}+{\sqrt[{5}]{7}}}}$:

$289*x^{60}-9460000*x^{54}-2392500000*x^{48}-190781250000*x^{42}+1127929687500*x^{36}-249084472656250*x^{30}+12817382812500000*x^{24}-457763671875000000*x^{18}+10728836059570312500*x^{12}-149011611938476562500*x^{6}+931322574615478515625$ (ok .. this is ugly)

Polynomial with root ${\sqrt[{4}]{{\sqrt {5}} \over {\sqrt[{3}]{{\sqrt {2}}+{\sqrt[{5}]{7}}}}}}$:

$289*x^{240}-9460000*x^{216}-2392500000*x^{192}-190781250000*x^{168}+1127929687500*x^{144}-249084472656250*x^{120}+12817382812500000*x^{96}-457763671875000000*x^{72}+10728836059570312500*x^{48}-149011611938476562500*x^{24}+931322574615478515625$ (very ugly .. especially considering you are not done yet and the next matrix to compute will be quite large)

This is how far I got so far. The program itself will do any polynomials just fine (I tweaked it a lot to get the final poly, so it may still have some bugs, however it works). The problem is working with such large numbers. The reducing part will multiply with 289 each time a $x^{240}$ is reduced, so we could potentially get coefficients of the order $289^{964-240}$, which is very large indeed. On my computer, I get about 30% done, when it starts running out of memory and I had to stop. This is why the program just generates the matrices for the last part (and by the way, Maxima crashed when trying to load the matrix to invert, so I will have to find another program that will handle such large sets of data)