Chinese Remainder Theorem, Rational numbers 27 October Theorem (Chinese remainder theorem). If {n1, . . . , nr } is a set of r natural numbers that are pairwise relatively prime, and if {a1, . . . , ar } are any r integers, then the system of congruences x ≡ a1 mod n1 ... x ≡ ar mod nr has a unique solution modulo N = Q ni. Here’s one way to find the solution. For each i, let N . ni Let yi be a solution to the congruence Ni = N i yi ≡ 1 mod ni. Set x= X j a j N j yj . Exercise 1. Find all solutions to the system of congruences n≡3 mod 4 n≡1 mod 21 n≡4 mod 5 Solution. Set N1 = 105 N2 = 20 N3 = 84. We must solve 105y1 ≡ 1 20y2 ≡ 1 84y3 ≡ 1 mod 4 mod 21 mod 5. For example y1 = 1, y2 = −1, y3 = −1 will work. Then a solution is x = (3)(105)(1) + (1)(20)(−1) + (4)(84)(−1) = 315 − 20 − 336 = −41. All solutions are of the form x = −41 + (4)(21)(5)k for k an integer. Rational numbers The problem of “defining” the rational numbers is the problem of how to “define” the fractions a/b. Let A be the set A = {(a, b) ∈ Z × Z|b 6= 0}. We suggestively write ab for the pair (a, b). The set A has “too many” rational numbers, because for example 1 2 6= . 2 4 We define a relation on R ⊂ A × A on A by a c ( , )∈R b d if ad = bc. Lemma 1. R is an equivalence relation. Definition 1. The set of rational numbers is the set of equivalence classes of A under the relation R. Proof. First, symmetry. Clearly a a ( , )∈R b b since ab = ba. Second, reflexivity. Suppose a c ( , )∈R b d i.e. ad = bc. Then cb = da so c a ( , ) ∈ R. d b Finally, transitivity. Suppose a c ( , )∈R b d and c e ( , )∈R d f i.e. ad = bc cf = de. We need to check that af = be. Since d 6= 0, we can check this after multipliying by d. adf = bde. This is bcf = bde, which is true. One checks that the formulae a c ad + bc + = b d bd ac ac = bd bd define addition and multiplication on Q, with the usual distributivity, associativity, and commutativity conditions. 0 0 a c c a Example 1. Show that if b = b0 and d = d0 , then a c a0 c0 + = 0+ 0 b d b d in Q. Proof. By definition c ad + bc a + = b d bd a0 c0 a0d0 + b0c0 + 0= . 0 0 0 b d bd The two right-hand-sides are the same if (ad + bc)(b0d0) = (a0d0 + b0c0)(bd) or adb0d0 + bcb0d0 = a0d0bd + b0c0bd (2) We have ab0 = a0b (3) cd0 = c0d. (4) Substituting these into the left side of (2) gives a0dbd0 + bc0b0d, which equals the right side of (2). 0 0 a a c c Exercise 2. Show that if b = b0 and d = d0 , then ac a 0 c0 = 0 0 bd b d in Q. Earlier we showed that if k has a rational square root, then it has an integral square root. More generally, we have Theorem . Suppose that f (x) = c0 + c1x + · · · + cnxn is a polynomial with integer coefficients. If r is a rational number such that f (r) = 0, and if r = pq in lowest terms, then p|c0 and q|cn. Proof. We have pn p 0 = c0 + c1 + · · · + cn n . q q Multiply through by q n to get 0 = c0q n + c1pq n−1 + · · · + cnpn. or −c0q n = c1pq n−1 + · · · + cnpn. Since p divides the right side, it divides the left. Since q is relatively prime to p, it follows that p divides c0. A similar argument shows q|cn. Example 2. Let k be an integer and n a natural number. If r is a rational number such that rn − k = 0 then r is an integer. Proof. Let f (x) = −k + xn so c0 = −k and cn = 1. If r = pq in lowest terms, then q divides 1 and so q = 1 and r is an integer. Exercise 3 (8.20). Let f (x) = x6 + cx5 + 1. Prove that the equation f (x) = 0 has no rational solution when c 6= ±2. Rational points on the circle. Another representation of rational numbers is the following. Let L be the set of lines L starting at the origin in R2 and pasing through some point (a, b) with a, b ∈ Z. There is a function m L −→ Q defined by m(L) = slope of L. Theorem . The function m is a bijection. Here is another instance of that idea. Theorem . The points (x, y) ∈ R such that x2 + y 2 = 1 are either of the form (−1, 0) or (x, y) = 1 − t2 2t , 1 + t2 1 + t2 ! for some real number t. The rational points are those for which t is rational. Proof. First observe that such these points are on the circle. Next, suppose that (x, y) is on the circle, and is not the point (−1, 0). The line through (−1, 0) and (x, y) is given by the (point-slope) formula Y = t(X + 1). This line intersects the circle at (x, y) if x2 + t2(x + 1)2 = 1 or (1 + t)2x2 + 2t2x + t2 − 1 = 0 or 2−1 2t t x2 + = 0. x+ (1 + t2)2 (1 + t2) One solution is x = −1, and so the other must be 1 − t2 . 2 1+t Note that x and y are rational if t is; and t = y/(x + 1) is rational if x is. Pythagorean triples A Pythagorean triple is three integers a, b, c such that a2 + b2 = c2. Theorem . The Pythagorean triples are integer multiples of triples of the form (2rs, r 2 − s2, r 2 + s2, r 2 − s2, 2rs, r 2 + s2), where r and s are integers. Exercise 4 (8.25). Give an example of a Pythagore triple which is not of the form (r 2 −s2, 2rs, r2 + s2).