Congruences - Algebra, CSIR-NET Mathematical Sciences | Mathematics for IIT JAM, GATE, CSIR NET, UGC NET PDF Download

Congruences are an important and useful tool for the study of divisibility. As we shall see, they are also critical in the art of cryptography.

Definition 3.1 If a and b are integers and n > 0,we write

a ≡ b mod n

to mean n|(b − a). We read this as "a is congruent to b modulo (or mod) n.

For example, 29 ≡ 8 mod 7, and 60 ≡ 0 mod 15.

The notation is used because the properties of congruence "≡" are very similar to the properties of equality"≡". The next few result make this clear.

Theorem 3.2 For any integers a and b, and positive integer n, we have:

1. a ≡ a mod n.
2. If a ≡ b mod n then b ≡ a mod n.
3. If a ≡ b mod n and b ≡ c mod n then a ≡ c mod n

These results are classically called: 1. Reflexivity; 2. Symmetry; and 3. Transitivity. The proofisasfollows:

1. n|(a − a) since 0 is divisible by any integer. Therefore a ≡ a mod n. b

2. If a ≡ b mod n then n|(b − a). Therefore, n|(−1)(b − a)or n|(a − b). Therefore,

3. If a ≡ b mod n and b ≡ c mod n then n|(b − a)and n|(c − b). Using the linear combination theorem, we have n|(b − a + c − b)or n|(c − a). Thus, a ≡ c mod n.

The following result gives an equivalent way of looking at congruence. It replaces the congruence sign with an equality.

Theorem 3.3 If a ≡ b mod n then b = a + nq for some integer q, and conversely.

Proof: If. a ≡ b mod n then by definition n|(b − a). Therefore, b − a = nq for some q.Thus then b = a + nq. Conversely if b = a + nq, then b - a = nq and so n|(b - a) and hence a ≡ b mod n then b = a + nq.

We will use often this theorem for calculations. Thus, we can write 15 ≡ −2 mod 17 by subtracting 17 from 15: −2=15 + (−1) .17. Similarly, 52 ≡ 12 mod 20. Just subtract 40 (2 times 20) from 52.

A simple consequence is this: Any number is congruent mod n to its remainder when divided by n. For if a = nq + r, the above result shows that a ≡ r mod n. Thus for example, 23 ≡ 2 mod 7 and 103 ≡ 3 mod 10. For this reason, the remainder of a number a when divided by n is called a mod n. In EXCEL, as in many spreadsheets, this is written ”MOD(a,n).” If you put the expression =MOD(23,7) in a cell, the readout will be simply 2. Try it!

Another way of relating congruence to remainders is as follows.

Theorem 3.4 If a ≡ b mod n then a and b leave the same remainder when divided by n. Conversely if a and b leave the same remainder when divided by n, then a ≡ b mod n.

Proof: Suppose a ≡ b mod n. Then by Theorem 3.3, b = a + nq. If a leaves the remainder r when divided by n, we have a = nQ + r with 0 < r < n. Therefore, b = a + nq = nQ + r + nq = n(Q + r) + r, and so b leaves the same remainder when divided by n. The converse is straightforward and we omit the proof.

We can now show some useful algebraic properties of congruences. Briefly, congruences can be added and multiplied.

Theorem 3.5 If a ≡ b mod n and c ≡ d mod n then
1. a + c = b + d mod n.
2. ac = bd mod n.

Proof: Write b = a + nq₁ and d = c + nq₂, using Theorem 3.3. Then adding equalities, we get b + d = a + c + nq₁ + nq₂ = a + c + n(q₁ + q₂). This shows that a + c ≡ b + d mod n by Theorem 3.3.

Similarly, multiplying, we get bd = (a + nq₁)(c + nq₂) = ac + naq₂ + ncq₁ + n²q₁ q₂. Thus, bd = ac + n(aq₂ + cq₁ + nq₁q₂, and so ac ≡ bd mod n, again by Theorem 3.3.

Some Examples.

We have noted that 23 ≡ 2 mod 7. We can square this (i.e. multiply this congruence by itself) to get 23² ≡ 4 mod 7. What a nice way to find the remainder of 23² when it is divided by 7! Multiply again by 23 ≡ 2 mod 7, to get

23³ ≡ 8 ≡ 1 mod 7

(This string of congruences is similar to a string of inequalities. It is read 23³ is congruent to 8 which is congruent to 1 mod 7. By transitivity (Theorem 3.2) this implies that 23³ is congruent to 1 mod 7.) Once we know that 23³ ≡ 1 mod 7, we can raise to the 5th powei (i.e. multiply this by itself 5 times) to get 23¹⁵ ≡ 1 mod 7. The application of a few theorems and we have found remainders of huge numbers rather easily!

Example 3.6 Find 17³⁴¹ mod 5. As explained on page 26, this is the remainder when 17³⁴¹ is divided by 5.

Method. We have

17 = 2 mod 5

Squaring, we have

17² ≡ 4 ≡ -1 mod 5

Squaring again, we find

17⁴ ≡ 1 mod 5

Now, 1 to any power is 1, so we raise this last congruence to the 85th power. Why 85? Just wait a moment to find out. We then find

17³⁴⁰  ≡ 1 mod 5

Finally, multiply by the first congruence to obtain

17³⁴1 ≡ 2 mod 5

So the required remainder is 2.

The strategy is to find some power of 17 to be 1 mod 5. Here, the power 4 worked. The we divided 4 into 341 to get a quotient 85, and this is the power we used on the congruence 17⁴ ≡ 1 mod 5. Note also the little trick of replacing 4 by -1 mod 5. This gives an easier number to square.

Example 3.7 Solve for x :5x ≡ 1mod 12.

One method is as follows. We know that gcd(5,12) = 1, so some linear combination of 5 and 12 is equal to 1. In Section 1 we had a general method for doing this, and we also had a spreadsheet approach. However, we can simply note by observation that

1 = 5 • 5+ (-2) • 12

So both sides of this equality are congruent to each other mod 12. Hence

1 = 5 • 5 +(-2) • 12 = 5 • 5 mod 12

So one solution is x = 5. More generally, if x ≡ 5 mod 12 then

5x ≡ 25 ≡ 1mod 12

Here is another approach: Start with the equation 5x ≡ 1 mod 12. If this were an equality, we would simply divide by 5 to get x = 1/5. But we are in the realm of integers so this won’t work. Instead we multiply by 5 to get 25x ≡ 5 mod 12 or x ≡ 5 mod 12. Note that we multiplied by 5 to get a coefficient of 1: 5 . 5 ≡ 1 mod 12.

The algebra of congruences is sometime referred to as “clock arithmetic.” This example illustrates this. Imagine you are a mouse and that each day you travel clockwise around a clock, passing through 25 minutes on the clock. You start at 12 o’clock. Here is what you journey will look like:

Note that the transition from 10 o’clock was not to 15 o’clock, but (working mod 12) to 15 mod 12 or 3 o’clock. In terms of clocks, we asked when the mouse would land at the 1 o’clock spot on the clock.

We can quickly nd when the mouse will land at 4 o‘clock. The equation is

5x ≡ 4mod 12

Multiply by 5 to get 25x ≡ 20 mod 12 or simply x ≡ 8 mod 12. It take 8 days.

Example 3.8 Same clock, different mouse. This mouse goes 23 minutes a day and starts at 12 o'clock. How many days before she reaches 9 minutes before 12?

The appropriate congruence is 23x ≡ −9 mod 60. We’ll use the gcd method and nd 1 as a linear combination of 23 and 60. A spreadsheet calculation gives

1= −13 . 23 + 5 .60

Taking this mod 60, we find

23(−13) ≡ 1mod 60.

Multiply by −9to get

23(117) ≡−9mod 60:

But 117 ≡ 57 mod 60. And so the mouse must travel 57 days to reach 9 minutes before the hour. Note that 57 ≡−3 mod 60 so the mouse will take 3 days if she goes in the other direction.

Up to now, all of our congruences have been modulo one fixed n. The following results show how to change the modulus in certain situations.

Theorem 3.9 If a ≡ b mod n, and c is a positive integer, then ca ≡ cb mod cn

Proof: This is little more than a divisibility theorem. Since n|(b — a), we have cn|c(b — a) or cn|(cb — ca), and this is the result.

The converse is also valid. Thus, if ca ≡ cb mod cn with c > 0 then a ≡ b mod n.

These results can be stated: A congruence can by multiplied through (including the modulus) and similarly, it can be divided by a common divisor.

Finally, we can mention that if a ≡ b mod n and if d|n, then a ≡ b mod d. We leave the proof to the reader.

We can now tackle the general question of solving a linear congruence ax ≡ b mod n. We will find when this congruence has a solution, and how many solutions it has. We first consider the case gcd(a,n) = 1. (In the examples above, this was the situation.) The following theorem answers this question and also shows how to find the solution.

Theorem 3.10 If gcd(a,n) = 1, then the congruence ax ≡ b mod n has a solution x = c. In this case, the general solution of the congruence is given by x ≡ c mod n.

Proof: Since a and n are relative prime, we can express 1 as a linear combination of them:

ar + ns =1

Multiply this by b to get abr + nbs = b. Take this mod n to get

abr + nbs ≡ b mod n or abr ≡ b mod n

Thus c = br is a solution of the congruence ax ≡ b mod n. In general, if x ≡ c mod n we have ax ≡ ac ≡ b mod n.

We now claim that any solution of ax ≡ b mod n is necessarily congruent to c mod n. For suppose ax ≡ b mod n. We already know that ac ≡ b mod n. Subtract to get
ax - ac ≡ 0 mod n or a(x - c) = 0 mod n

But this means that n|a(x - br). But since a and n are relatively prime, this implies that n|(x - c) and x ≡ c mod n. This completes the proof.

An important special case occurs when n is a prime p.

Corollary 3.11 If p is a prime, the congruence ax ≡ b mod p has a unique solution x mod p provided  mod p.

The reason we single this case out is that this result is almost exactly like the similar result in high school algebra: The equation ax = b has a unique solution provided a ≠ 0. We shall soon delve further into this analogy. The reason this is true is that if an integer a is not divisible by p, it is relatively prime to p. Thus, if  mod p, then a and p are relatively prime.

During the course of the proof of theorem 3.10 , we proved the following useful result.

Theorem 3.12 If ab ≡ ac mod n and if gcd(a, n) = 1, then we have b ≡ c mod n.

In short, we can cancel the factor a from both sides of the congruence so long as gcd(a, n) = 1. In algebra, we learn that we “can divide an equation ax = ay by a” if a ≠ 0. Here we can “cancel the factor a from both sides of the congruence ax ≡ ay mod n” if a and n are relatively prime. This theorem is sometimes called the cancelation law for congruences.

Now suppose that we wish to solve the congruence ax ≡ b mod n where d = gcd(a, n) > 1. For example, consider the congruence 18x ≡ 12 mod 24. Here d = gcd(18, 24) = 6. We can divide this congruence by 6 to get the equivalent15 congruence 3x ≡ 2 mod 4. So we end up with the congruence 3x ≡ 2 mod 4, in which gcd(3, 4) = 1 and which has general solution x ≡ 2 mod 4. So this is the solution of the original congruence 18x ≡ 12 mod 24. This worked because the gcd also divided the constant term 12. If it didn’t there would be no solution. This is the content of the following theorem which generalizes this problem.

Theorem 3.13 Given the congruence ax ≡ b mod n. Let d = gcd(a,n). Then

1. If d does not divide b, the congruence has no solution.
2. If d\b then the congruence is equivalent to the congruence (a/d)x ≡ (b/d) mod (n/d) vjhich has a unique solution mod n/d.

Proof: Suppose there were a solution of ax ≡ b mod n. Then we would have ax ≡ b mod d. But a ≡ 0 mod d since d|a. So we would have 0 ≡ b mod d or d|b. So a necessary condition for a solution is that d|b. This prove part 1. As for part 2, divide the entire congruence by d as in the above example. The reduced congruence has a unique solution mod n/d since a/d and n/d are relatively prime.

Algebra on a Small Scale.

Corollary 3.11 has an interesting interpretation-if p is a prime and we work mod p, the integers mod p behave algebraically like the real numbers. In the real number system the equation ax = b has a solution x = b/a = ba^-1 where a^-1 = 1/a is the reciprocal of a and is the solution of the equation ax = 1. What is the situation if we try to do this mod p?

Example 3.14 What is the value of 5⁻¹ mod 7?

Method. It is required to nd the solution of 5x ≡ 1 mod 7. We can do this using the method of Example 3. Since

3 .5+(−2)7 = 1

be observation, we have

3 . 5 ≡ 1mod 7

So 5⁻¹ ≡ 3 mod 7, or simply 5⁻¹ = 3 mod 7, where equality if used because it is understood that we are working mod 7.

Since we are working mod 7, there are only 7 different numbers mod 7, namely the remainders 0 through 6 when a number is divided by 7. So the algebra of numbers mod 7 is a strictly nite algebra. Here is the multiplication table for these numbers mod 7. We omit 0.

The number 1 is underlined in the body of the table. The row and column where a 1 appears are inverses, because the product is 1. By observation, we can see that 2 and 4 are inverses mod 7, as are 3 and 5. Both 1 and 6 are self inverses. (Note that 6 = −1 mod 7, and so it is not surprising that 6 is its own inverse: (−1)⁻¹ = −1.

Let us go one step further with the analogy with ordinary algebra.

Example 3.15 Solve the congruence 8x ≡ 13 mod 29.

First method. In analogy with algebra we expect the solution x ≡ 13 .8⁻¹ mod 29. So we rst compute 8⁻¹ mod 29. We express 1 as a linear combination of 8 and 29 by the method given in section 1, or using a spreadsheet. A possible result is

1= 11 .8 − 3. 29

Taking this mod 29, we nd 8−1 ≡11 mod 29. So, solving for x, we find

x ≡ 13 . 8−1 ≡ 13 .11 = 143 ≡ 27 mod 29

Second method. Using fractions, we write

Ordinarily, we cancel factors in the numerator and denominator. We can’t do this here, but we can multiply numerator and denominator by the same (non-zero) number. We choose 4, because this gets the denominator close to the modulus 29, making the quotient simpler.
Thus

Now do it again, using a factor 10:

This is the same answer, of course. Here’s the way the full solution works in one line:

Third method. When we write  mod 29, we can cancel at least one factor 2, if we add 29 to the numerator. Thus,

We don’t necessarily recommend this method, but we use it to illustrate that there are often many ways to attack a problem and to show the inner consistency of our small scale arithmetic.

Divisibility Tricks. The number 345,546,711 is divisible by 3. In fact it is divisible by 9. We can discover this easily using the following trick, which we shall prove.

A number is congruent mod 9 to the sum of the digits in that number.

Here we have

345; 546; 711 ≡ 3+4+5 + 5+4+6+7+1+1 = 36 ≡ 3+6 = 9 ≡ 0 mod 9

In fact, using this result, it is not even necessary to nd the sum. There are short cuts. For example 3 + 4 + 5 = 12 which is congruent to its digit sum 1 + 2 = 3 mod 9. Continuing, add 5 + 5 = 10 ≡ 1, so we add 1 to 3 to get 4. And so on. This is a lot easier to do than to explain. Briefly, any time you get a two digit answer, replace it by its digit sum.

The proof of this trick depends on the knowledge that the digits in an expansion of a number represent coecient of powers of 10. Thus,

3;412 = 3 x10³ +4x 10² +1 x 10¹ +2 x1

Since 10 ≡ 1 mod 9, we can square to get 10² ≡ 1 mod 9. Similarly, by cubing we get 10³ ≡ 1 mod 9, and so on. Thus,

3412 = 3 x 10³ + 4 x 10² + 1 x 10¹ + 2 x 1 ≡ 3 + 4+1 + 2 mod 9

where the latter sum is simply the sum of the digits of 3412. This generalizes to give the result. It follows that a number is congruent to its digit sum mod 3, because if a ≡ b mod n and d|n then a ≡ b mod n.(Here n = 9 and d = 3.)

This simple trick has a useful application. It is a check on possible calculation errors. For example, suppose you are given the multiplication 341 x 167 = 56847 and you are suspicious of this result. (Perhaps someone was sloppy or didn’t copy it down correctly.) Now if this multiplication were true, it would also be true mod 9. But 341 ≡ 8 mod 9 (just add the digits!) and 167 ≡ 14 ≡ 5 mod 9 so 341 x 167 ≡ 8 x 5 = 40 ≡ 4 mod 9. But the answer given us was 56847 ≡ 30 ≡ 3 mod 9, and so it was in error. This method is not failsafe, but it is a quick check.¹⁶ Incidentally, you know that the multiplication 1234567 x 245678 = 303305951435 is wrong. (Hint: look at the last digits.) You know it’s wrong by checking the answer mod 10.

There is another simple trick to find a number mod 11 using its digits. In this case, we find the alternating sum starting with the units column. For example, to find 56744 mod 11, we compute 56743 = 3 - 4 + 7 - 6 + 5 = 5 mod 11. The proof is similar to the proof above, and is based on the simple congruence 10 = ≡ -1 mod 11. Squaring, we get 100 =≡ 1 mod 11. Cubing, we get 1000 ≡ 1 mod 11, etc. Thus,

56743 = 3 + 4 x 10 + 7 x 10² + 6 x 10³ + 5 x 10⁴ ≡ 3 - 4 + 7 - 6 + 5 = 5 mod 11

The general proof is the same.

For example, the alleged calculation 345 x 3456 = 1129320 can be check mod 11. We have

345 x 3456 = (5 - 4 + 3)(6 - 5 + 4 - 3) = 4 x 2 = 8 mod 11

The alleged answer is 1129320 = 0 - 2 + 3- 9 + 2- 1 + 1 = -6 ≡ 5  8 mod 11. The actual answer for this multiplication is 1192320. so the error was a simple transposition of digits, a common error. The alternating sum will catch such an error.