MHB System of congruences, not relatively prime moduli

  • Thread starter Thread starter evinda
  • Start date Start date
  • Tags Tags
    Prime System
Click For Summary
The discussion focuses on solving a system of congruences where the moduli are not relatively prime. The initial equations are transformed to analyze their individual components, leading to the realization that the Chinese Remainder Theorem (CRT) cannot be applied directly due to the lack of co-primality. Instead, an intermediate approach is suggested, involving finding the greatest common divisor (gcd) to establish the existence of solutions. A specific solution is identified, and the general solution is derived using the least common multiple (lcm) of the moduli. The conversation emphasizes the importance of checking conditions for the existence of solutions when dealing with non-coprime moduli.
evinda
Gold Member
MHB
Messages
3,741
Reaction score
0
Hello! (Wave)

I want to solve the following system of congruences:

$$x \equiv 13 \pmod{40} \\ x\equiv 5 \pmod{44} \\ x \equiv 38 \pmod{275}.$$I have thought the following:

$$x \equiv 13 \pmod{40} \Leftrightarrow x \equiv 13 \pmod{2^3 \cdot 5}$$

$$x \equiv 5 \pmod{44} \Leftrightarrow x \equiv 5 \pmod{2^2 \cdot 11}$$

$$x \equiv 38 \pmod{275} \Leftrightarrow x \equiv 38 \pmod{5^2 \cdot 11}$$$$x \equiv 13 \pmod{2^3 \cdot 5} \Leftrightarrow x \equiv 13 \pmod{2^3} \text{ and } x \equiv 13 \pmod{5} \ \ (1)$$

$$x \equiv 5 \pmod{2^2 \cdot 11} \Leftrightarrow x \equiv 5 \pmod{2^2} \text{ and } x \equiv 5 \pmod{11} \ \ (2)$$

$$x \equiv 38 \pmod{5^2 \cdot 11} \Leftrightarrow x \equiv 38 \mod{5^2} \text{ and } x \equiv 38 \pmod{11} \ \ (3)$$

$(1)$: $x \equiv 5 \pmod{2^3}$ and $x \equiv 3 \pmod{5}$

$(2)$: $x \equiv 1 \pmod{2^2}$ and $x \equiv 5 \pmod{11}$

$(3)$: $x \equiv 13 \pmod{5^2}$ and $x \equiv 5 \pmod{11}$Am I right so far?

How can we continue? Can we somehow apply the Chinese Remainder Theorem? (Thinking)
 
Mathematics news on Phys.org
Hey evinda!

Unfortunately we cannot apply CRT directly since the modulo numbers are not co-prime.

We can solve the problem however, by working through the equations as follows:

$x\equiv 13 \pmod{40}\Rightarrow x=13+40k \tag 1$
$x \equiv 5 \pmod{44} \Rightarrow 13+40k \equiv 5 \pmod{44} \Rightarrow 40k \equiv -8 \pmod{44} \tag 2$

Normally we can solve (2) directly by multiplying with the inverse of $40$ with respect to $44$, but in this case this inverse doesn't exist because $40$ and $44$ are not co-prime.
So instead we make an intermediate step, and then get the inverse:
$$40k = -8 + 44\ell \Rightarrow 10k=-2+11\ell \Rightarrow 10k\equiv -2 \pmod{11} \\ \Rightarrow k \equiv [10]^{-1}_{11} \cdot -2 \pmod{11} \Rightarrow k= [10]^{-1}_{11}\cdot -2 + 11m$$
where $[10]^{-1}_{11}$ is the inverse of $10$ modulo $11$.

Can we find $k$ now? And substitute it back into (1)?
Afterwards, we can repeat with the last equation. (Thinking)
 
Some time ago, I made an observation on the S.O.S. forum on what happens with simultaneous congruence equations when the modulo numbers are not coprime: Chinese remainder theorem.

Let’s take the first two equations: $x\equiv13\pmod{40}\equiv5\pmod{44}$. A solution exists if and only if $13\equiv4\pmod d$ where $d=\gcd(40,44)=4$. This holds, and so a solution exists. Noting that $x=93$ satisfies the congruences and $\mathrm{lcm}(40,44)=440$, the general solution of the congruence is $x=93+440k$, $k\in\mathbb Z$.

Now we do the same for the congruences $x\equiv93\pmod{440}\equiv38\pmod{275}$. First check that a solution exists: $\gcd(440,275)=55$, $93\equiv38\pmod{55}$. It does. Then find $l=\mathrm{lcm}(440,275)$, find a particular solution $x_0$ to the congruences, and the general solution will be of the form $x=x_0+lk$, $k\in\mathbb Z$.

$x=1413+2200k$, $k\in\mathbb Z$.
 
Seemingly by some mathematical coincidence, a hexagon of sides 2,2,7,7, 11, and 11 can be inscribed in a circle of radius 7. The other day I saw a math problem on line, which they said came from a Polish Olympiad, where you compute the length x of the 3rd side which is the same as the radius, so that the sides of length 2,x, and 11 are inscribed on the arc of a semi-circle. The law of cosines applied twice gives the answer for x of exactly 7, but the arithmetic is so complex that the...