Conjecture: $xa^y+yb^z=zc^w+wd^x$ has no solutions in distinct positive integers satisfying $a + b + c + d = x + y + z + w$
问题内容
Motivation:
While recently researching equations similar to this, I suddenly had an idea: would adding coefficients and specific restrictions change the difficulty? Therefore, I unexpectedly pieced together the following Diophantine equation:
Let $a, b, c, d, x, y, z, w$ be 8 pairwise distinct positive integers, satisfying:
$$a + b + c + d = x + y + z + w$$
Conjecture: The following exponential equation has no positive integer solutions:
$$x \cdot a^y + y \cdot b^z = z \cdot c^w + w \cdot d^x$$
I wrote a pythoncompiler to search for counterexamples. No valid 8-tuple can be found within the range $1 \le a, b, c, d, x, y, z, w \le 30$. (This is already the limit of my computer :D)
My question:
Can this conjecture be proven using existing theorems? Or are there counterexamples with larger values? Or are there any further restrictions on $a, b, c, d, x, y, z, w$ (for example, that they are all odd numbers)? Any answers are welcome!
Update (Based on @User's answer)
@User's counterexample has refuted the original conjecture.
As he suggested in his answer, if we further restrict all 8 variables to pairwise distinct positive odd numbers:
$$\text{Does } x \cdot a^y + y \cdot b^z = z \cdot c^w + w \cdot d^x \text{ have any solutions when } a,b,c,d,x,y,z,w \text{ are all odd?}$$
Since for odd numbers $u^n \equiv u \pmod 8$, the modulo $8$ operation does not rule out the existence of a solution. I welcome any insights into this case containing only odd numbers!
Update 2
Since @User demonstrated that an infinite parametric family of counterexamples exists even under the all-odd constraint.
Refined Question:
Let $a, b, c, d, x, y, z, w$ be 8 pairwise distinct PRIME numbers satisfying:
$$a + b + c + d = x + y + z + w$$
Does the Diophantine equation
$$x \cdot a^y + y \cdot b^z = z \cdot c^w + w \cdot d^x$$ have any finite solutions?
回答 (1)
I managed to find a counterexample; I will provide all the calculations below, but the conjecture is false $$(a,b,c,d,x,y,z,w)=(193,1,6,1007,2,3,1197,5)$$
Let $b=1$, $x=2$, $y=3$, $c=6$, $w=5$, then it follows from the condition that $z=a+d-3$, that is, all that remains is to solve $5d^2+7776d-12,900,677=0$. Notice that when $a=193$, the equation becomes the quadratic equation $5d^2+7776d-12,900,677=0$, hence, $d=1007$ and $z=193+1007-3=1197$ $$193+1+6+1007=1207=2+3+1197+5$$
\begin{aligned} 2\cdot193^3+3\cdot1^{1197} &=14\,378\,117\\ 1197\cdot6^5+5\cdot1007^2 &=9\,307\,872+5\,070\,245\\ &=14\,378\,117 \end{aligned}
If we require all eight numbers to be odd, then the counterexample no longer applies — this is a separate strengthened conjecture. There is no simple contradiction from parity, that is, for odd $u$, $n$ we have $u^n\equiv u\bmod 8$, hence, $xa+yb\equiv zc+wd\bmod 8$
The question was about odd numbers
Let $\displaystyle T=\frac{3^{63}+3^{21}+3^9+3^7}{20}$, then
$$(a,b,c,d,x,y,z,w)=(3^{63},3^9,3^{21},3^7,9T,T,7T,3T)$$
The condition on the sums is satisfied
$$a+b+c+d=3^{63}+3^9+3^{21}+3^7=20T$$
$$x+y+z+w=9T+T+7T+3T=20T,\quad a^y=b^z=c^w=d^x=3^{63T}$$
Let $P$ be the common value, then $xa^y+yb^z=(9T+T)P=10TP$ and $zc^w+wd^x=(7T+3T)P=10TP$, so the original exponential equality is also satisfied. Moreover, instead of $3$, one can take any odd $r\equiv3,5,7\bmod {10}$ and define $\displaystyle T_r=\frac{r^{63}+r^{21}+r^9+r^7}{20}$ this will be an odd integer, and $(r^{63},r^9,r^{21},r^7,9T_r,T_r,7T_r,3T_r)$ will again give a solution