December 9, 2004 Lecturer: Dr Martin Kurth Michaelmas Term 2004 Problem Sheet 3 – Solutions 1. The function f approaches the limit lim f (x) = L x→x0 if for every real number ² > 0 there is a real number δ > 0 such that for all x in the domain of f 0 < |x − x0 | < δ ⇒ |f (x) − L| < ². 2. For x 6= 3, the first function can be rewritten as f (x) = (x + 3)(x − 3) = x + 3, x−3 and thus lim f (x) = lim (x + 3) = 3 + 3 = 6. x→3 x→3 The second and third functions haven’t got a limit as they “grow too large” as x approaches 3. The fourth function “jumps” at x = 3, and thus has no limit. 3. This is more complicated. The answer is that the function f does not have a limit at x=0. To prove this, we first prove some lemmas (auxiliary theorems). • Lemma 1: If an integer number m is even, m2 is even. Proof: As m is even, there exists an integer number k with m = 2k, and thus m2 = 4k 2 , and as k 2 is an integer number, m2 is even. • Lemma 2: If an integer number n is odd, n2 is odd. Proof: As n is odd, n + 1 is even, so we know from Lemma 1 that (n + 1)2 is even. Now we have (n + 1)2 = n2 + 2n + 1, and thus n2 = (n + 1)2 − 2n + 1. As we know that (n + 1)2 is even, and 2n is even by definition, (n + 1)2 − 2n is even, so n2 must be odd. 1 √ • Lemma 3: 2 is an irrational number (ie not a√rational number). Proof (by contradiction): Let us assume that 2 is rational. Then √ 2 can be written as √ m 2= n with m and n integer numbers that haven’t got a common divisor (if they have, just cancel the common divisor, so we end up with m and n that haven’t got one). Now we write √ m2 ( 2)2 = 2 = 2 , n ie m2 = 2n2 . This means that m2 is even, and thus m must be even (if it was odd, then m2 would be odd by Lemma 2), so there exists an integer number k, such that m = 2k, which means m2 = 4k 2 . This gives us 4k 2 = 2n2 , which means n2 = 2k 2 , ie n2 is even, and with the same argument as for m we find out that n is even. We have now shown that m and n are even, ie they have 2 as a common divisor, contrary to the assumption that they have √ no 2 is common divisor at all. This means that the assumption that √ rational is wrong, hence 2 must be irrational. • Lemma 4: For every real number δ > 0 there exists a rational number x with |x| < δ. Proof: Choose an integer number q with q> 1 . δ x= 1 q Then is rational and satisfies |x| = as q > 1/δ. 2 1 <δ q • Lemma 5: For every real number δ > 0 there exists an irrational number z with |z| < δ. Proof: From Lemma 4 we know that there is a rational number x of the form x = 1/q with integer q and |x| < δ. Now consider 1 z= √ . q 2 First we have to show that z is irrational. We do this by contradiction, so let us assume z is rational, ie there are integer numbers m and n with m z= . n This immediately gives us √ 2= n , mq √ √ ie 2 is rational. As we know from Lemma 3 that 2 is not rational, our √ assumption must be wrong, and z must be irrational. As 2 > 1 we also know that 1 1 |z| = √ < = |x| < δ, q q 2 which proves Lemma 5. Now we can proceed to prove our theorem that f has no limit at x = 0. Again, let us assume the opposite, ie that such a limit exists. • Assumption 1: lim = 1. x→0 Let us choose ² = 1/2. Now we must find a δ > 0, such that for all x 0 < |x| < δ ⇒ |f (x) − 1| < ². But from Lemma 5 we know that for every δ we choose, there will be a number z with |z| < δ that is irrational, ie f (z) = 0. In other words: |f (z) − 1| = |0 − 1| = 1 > ². This means that 1 can’t be the limit of f (x) at x = 0. 3 • Assumption 2: lim = L 6= 1. x→0 Now let us choose 1 |1 − L|. 2 Then, for every δ we choose, we know from Lemma 4 that there is a number x with |x| < δ which is rational, ie ²= f (x) = 1. We thus have |f (x) − L| = |1 − L| > ², which means that L 6= 1 can’t be the limit of f (x) at x = 0 either. This means that f (x) has no limit as x approaches 0. Note: Don’t believe that mathematicians think in this structured way, ie they sit down, and think “I’ll start with 5 lemmas, and Lemma 1 will be . . . ”. They will normally start to scribble down the proof of the “big” theorem, and then notice what lemmas they need as they go along. When they write their proof down properly, it looks as well planned as this page, but a mathematician’s scratch paper doesn’t look anything like this! 4