J’avoue ne jamais avoir réfléchi à la question! Que donne le nombre imaginaire i élevé à la puissance i? La réponse est surprenante mais c’est la démonstration de Matt Parker qui est encore plus intéressante!
Toutes les explications ici!
It all started this morning with a very short post in the newsgroup rec.puzzles :
This is a cryptarithm.
TRUTH * IS = WISDOM
Each letter is a different digit; T appears twice and represents the same digit, and so on; base 10.
I’ve solved puzzles similar to this one in the past but usually, I had a few hints or givens and it involved addition, not multiplication! So this one seemed more complex and interesting!
The method for solving this kind of puzzle is more or less always the same : convert arithmetic operations into algebra (or sets of possible values), eliminate a few candidate values by deduction, do some simple substitutions in the equations you get and voilà! But this time, I had to work a lot harder!
Let’s start with a graphic that details what goes on when multiplying the number TRUTH by the number IS to get the product, the number WISDOM.
The trick here is to list everything we know (or can deduct) about this product.
From the S * TRUTH
1) (S * H) = a + (10 * b)
2) (S * T) + b = c + (10 * d)
3) (S * U) + d = e + (10 * f)
4) (S * R) + f = g + (10 * h)
5) (S * T) + h = i + (10 * j)
From I * TRUTH
6) (I * H) = k + (10 * l)
7) (I * T) + l = m + (10 * n)
8) (I * U) + n = o + (10 * p)
9) (I * R) + p = q + (10 * r)
10) (I * T) + r = s + (10 * t)
From the intermediate results of Line 3 + Line 4
11) a + 0 = M + (10 * u), this is a + ZERO = M + (10 * u)
12) u + c + k = O + (10 * v)
13) v + e + m = D + (10 * w)
14) w + g + o = S + (10 * x)
15) x + i + q = I + (10 * y)
16) y + s + j + t = W
From the numbers themselves
We know that the 3 numbers cannot start with the digit zero so :
17) 1 <= W <= 9, or W <> 0
18) 1 <= I <= 9, or I <> 0
19) 1 <= T <= 9, or T <> 0
Since TRUTH is a 5-digit number, it has to be less than 99999. To be more precise, since all letter-digits have to be different, TRUTH would have to be less than the maximum possible (98765) but, because of the repeating digit T, we have:
20) 10000T + 1000R + 100U + 10T + H <= 98796
Same reasoning for IS :
21) (10*I) + S <= 98
Finally, for WISDOM
22) 100000W + (10000*I) + 1000S + 100D + (10*O) + D <= 987654
Same logic as above can now be applied to minimum values :
23) 10000T + 1000R + 100U + 10T + H >= 10213
24) (10*I) + S >= 10
25) 100000W + (10000*I) + 1000S + 100D + (10*O) + M >= 102345
One remark about H and S : if H or S would be zero, the resulting digit from this product, M, would also be zero. That’s a contradiction since all digits have to be different, therefore,
26) 1 <= H <= 9, or H <> 0
27) 1 <= S <= 9, or S <> 0
From the carry
Now, let’s have a look at the possible values for every carry. Let’s start with an example
Since all digits have to be different, the maximum result we can have from a multiplication in our case is 9×8 = 8×9 = 72, therefore the maximum value for all carry variables is 7. So,
28) b <= 7
29) d <= 7
30) f <= 7
31) h <= 7
32) j <= 7
33) l <= 7
34) n <= 7
35) p <= 7
36) r <= 7
37) t <= 7
For the addition carry (Carry of line 3 + line 4), since they only involve 2 digits, the maximum is 9+8 = 8+9 = 17, therefore the carry must be zero or one. So,
38) 0 <= u <= 1
39) 0 <= v <= 1
40) 0 <= w <= 1
41) 0 <= x <= 1
42) 0 <= y <= 1
Now, you can throw in a few modular equations and arithmetic, build sets of possible values and carefully examine W and use the fact that the multiplication I*T appears twice. Also take advantage of the fact that all equations are diophantine equations! It’s also very useful to notice the fact that T, H, S, I and W cannot be zero! I hope you have enough to get you started!
I hope you didn’t think I’d spoil all the fun, don’t you? The rest is left as an exercise to the reader as they say!
Oh, by the way, I verified and there’s only one solution! ;)
Fignolez vos papiers, corrigez vos publications : le call for paper de la conférence 11th Workshop on Implementation, Compilation, Optimization of OO Languages, Programs and Systems est annoncé ici!
Vous aimez les maths? Vous aimez Smalltalk? Didier H. Besset attend vos suggestions!
Pharo Days : les détails pour les inscriptions sont ici.
Visualiser les dépendances entre des packages FAMIX avec Moose/Pharo.
SqueakJS et ElectronJS : curieux?
Un cadran construit à partir d’expressions mathématiques particulières… Le genre de truc qui m’amuse et m’intéresse. L’article est ici.