We know that limx→2(2x+1)=5. We usually say that we can make f(x)=2x+1 as close to 5 as we want, as long as x is close enough to 2, without letting x=2. Checking x=1.9,1.99 and 1.999 gives f(x)=4.8,4.98 and 4.998, but three numbers don't tell us about every x near 2. To write a rule that does, we turn "as close as we want" and "close enough" into distances.
What "close enough" means for y=2x+1 near (2,5)
Text solution
Write each distance as an absolute valueThe distance from x to 2 is ∣x−2∣, and the distance from f(x) to 5 is ∣f(x)−5∣.
∣x−2∣<0.5⟺1.5<x<2.5
∣f(x)−5∣<1⟺4<f(x)<6
A smaller ε needs a new δFor ε=0.1, the old δ=0.5 lets in x=2.4, whose output is too far from 5. Solving again gives δ=0.05.
ε=0.1,δ=0.5:x=2.4,f(2.4)=5.8>5.1
ε=0.1:4.9<2x+1<5.1⇒1.95<x<2.05⇒δ=0.05
ε=0.1,δ=0.5:x=2.4,f(2.4)=5.8>5.1
ε=0.1:4.9<2x+1<5.1⇒1.95<x<2.05⇒δ=0.05
One rule for every εThe slope of y=2x+1 is 2, so y changes twice as fast as x. To keep f(x) within ε of 5,x has to stay within 2ε of 2.
δ=2ε
Here's the picture for ε=1. The outputs from 4 to 6 carry across to the line and down to the inputs from 1.5 to 2.5, which are 0.5 from 2 on each side.
ε=1:4<2x+1<6⇒1.5<x<2.5⇒δ=0.5
Replacing 2 with a and 5 with L gives the precise definition of a limit. For every ε>0, there is a δ>0 such that all x within δ of a, except a itself, have f(x) within ε of L.
x→alimf(x)=L:
for every ε>0 there is a δ>0 such that
for every ε>0
there is a δ>0 such that
a=2,L=5:0<∣x−2∣<δ⇒∣(2x+1)−5∣<ε,δ=2ε
a=2,L=5:0<∣x−2∣<δ⇒∣(2x+1)−5∣<εδ=2ε
A limit never depends on the value at a itself, so we write 0<∣x−a∣ to say that x=a.
2
When the two sides give different distances
On a curve, finding δ has a catch. Take limx→9x=3 with ε=1, so x has to stay between 2 and 4. Carrying y=2 and y=4 across to the curve and down to the x-axis lands at x=4 and x=16.
x=2x=49−416−9⇒x=4⇒x=16=5=7
On the line, the distances to the left and right were equal. Here they are 5 and 7, but ∣x−9∣<δ keeps x between 9−δ and 9+δ, the same distance on both sides. Should we choose δ=5 or δ=7?
Find δ for limx→9x=3 when ε=1
x
y
y=x
2
3
4
ε=1
9
Text solution
δ=7 is too largeWith δ=7,x can go down to 2, and x=2.25 gives an output below 2.
δ=7:2<x<16,2.25=1.5<2
δ=7:2<x<162.25=1.5<2
δ=5 worksWith δ=5, every output stays between 2 and 14, which is less than 4.
δ=5:4<x<14⇒2<x<14<4
δ=5:4<x<14⇒2<x<14<4
Why the left side is tighterx has a larger slope to the left of 9, so it drops to 2 after only 5 units, while it takes 7 units on the right to reach 4. Any smaller δ, like 1, works too, and 5 is the largest that does.
3
Showing a limit does not exist
The definition can also show that a limit doesn't exist. From the left of 0, this f(x) is 1, and from the right, it is 3. The number 2 is right in the middle, so could 2 be the limit?
f(x)={1,3,x<0x≥0x→0limf(x)=2?
To show that 2 is not the limit, we only need one ε for which no δ works. With ε=21, the outputs would have to stay between 1.5 and 2.5.
x
y
0
1
3
2
1.5
2.5
ε=21:
∣1−2∣∣3−2∣=1≥21=1≥21
every δ>0:
f(−2δ)f(2δ)=1=3
Show with ε and δ that the limit at 0 is not 2, and that no limit exists
x
y
0
1
3
2
f(x)={1,3,x<0x≥0
Text solution
One generous ε proves nothingFor ε=1.5, both values are already inside, so every δ works. But the definition needs a δ for every positive ε.
ε=1.5:0.5<1<3.5,0.5<3<3.5
ε=1.5:0.5<1<3.5,0.5<3<3.5
No δ works for ε=21However small δ is, −2δ and 2δ are within δ of 0, and they give 1 and 3, both 1 away from 2. So 2 is not the limit, and any ε≤1 fails the same way.
No other L works eitherWith ε=21, the outputs must fit between L−21 and L+21, a range of width 1. It can't hold both 1 and 3, so limx→0f(x) does not exist.
∣3−1∣=2>1=2⋅21
Now try it yourself. Can you keep the outputs as close to the proposed limit as you choose?
Choosing epsilon and finding delta
4
Writing an ε-δ proof
On y=2x+1, we found δ=2ε by testing ε=1 and ε=0.1. A proof can't test numbers one by one. It has to give a δ for every positive ε, so we need a formula for δ in terms of ε.
One fact we'll use: ∣ab∣=∣a∣∣b∣. So when we factor a number out of an absolute value, it comes out as its absolute value, and both 2 and −2 come out as 2.
∣2x−4∣=∣2∣∣x−2∣=2∣x−2∣
∣6−2x∣=∣−2∣∣x−3∣=2∣x−3∣
Prove that limx→2(2x+1)=5 and limx→2x−2x2−4=4
x→2lim(2x+1)=5x→2limx−2x2−4=4:δ=?:δ=?
Text solution
Scratch work runs backwardsStart from ∣(2x+1)−5∣<ε and rewrite it until it's a condition on ∣x−2∣.
∣(2x+1)−5∣=∣2x−4∣=2∣x−2∣
2∣x−2∣<ε⟺∣x−2∣<2ε⇒δ=2ε
2∣x−2∣<ε⟺∣x−2∣<2ε⇒δ=2ε
The fraction: x=2 lets us cancelThe definition assumes 0<∣x−2∣, so x is never 2, and the fraction is x+2. Then ∣(x+2)−4∣ is just ∣x−2∣, so δ=ε works.
x−2x2−4−4=∣(x+2)−4∣=∣x−2∣<δ⇒δ=ε
x−2x2−4−4=∣(x+2)−4∣=∣x−2∣<δ⇒δ=ε
Here's how the proof for 2x+1 is written up. It runs forwards, the way the definition reads: choose δ, assume 0<∣x−2∣<δ, and end at ∣(2x+1)−5∣<ε.
Let ε>0. Choose δ=2ε.
If 0<∣x−2∣<δ, then ∣(2x+1)−5∣=2∣x−2∣<2δ=ε.
If 0<∣x−2∣<δ, then∣(2x+1)−5∣=2∣x−2∣<2δ=ε.
On a test, you only hand in the proof. The scratch work is how you found δ.
5
When the factor changes with x
For limx→3x2=9, the scratch work starts the same way. By the difference of squares, ∣x2−9∣ splits into ∣x−3∣ times another factor.
x2−9=(x−3)(x+3)
∣x2−9∣=∣x−3∣∣x+3∣
For 2x+1, the factor next to ∣x−2∣ was the fixed number 2. Here ∣x+3∣ changes with x, and δ has to be one number chosen before we know x, so we can't divide ε by it.
Prove that limx→3x2=9 using the ε-δ definition
x→3limx2=9:δ=?
Text solution
Why 6ε isn't enoughAt x=3 the factor is 6, but to the right of 3 it's larger. At the edge x=3+δ,∣x2−9∣ is already more than ε, so it is at some points just inside too.
δ=6ε,x=3+δ:∣x2−9∣=δ(6+δ)>6δ=ε
δ=6ε,x=3+δ:∣x2−9∣=δ(6+δ)>6δ=ε
Cap δ at 1 to bound the factorδ is ours to choose, and a smaller δ still works, so we decide δ≤1. That keeps ∣x+3∣ below 7. A cap of 2 would work as well, with 8 in place of 7.
δ≤1:2<x<4⇒5<x+3<7⇒∣x+3∣<7
δ≤1:2<x<4⇒5<x+3<7⇒∣x+3∣<7
Two jobs, so take the smallerδ≤1 keeps ∣x+3∣<7, and δ≤7ε makes 7δ≤ε. Without the cap, ε=14 would give δ=2, and x=4.9 gives 4.92−9=15.01>14.
δ=min(1,7ε)
The proofLet ε>0 and choose that δ. If 0<∣x−3∣<δ, then ∣x−3∣<1, so ∣x+3∣<7, and
∣x2−9∣=∣x−3∣∣x+3∣<7∣x−3∣<7δ≤ε.
∣x2−9∣=∣x−3∣∣x+3∣<7∣x−3∣<7δ≤ε.
Why do we need both 1 and 7ε? Change ε and see which one sets δ.
Two restrictions on delta
Summary
The definition of a limit
for every ε>0 there is a δ>0 such that
for every ε>0there is a δ>0 such that
0<∣x−a∣<δ⇒∣f(x)−L∣<ε
0<∣x−a∣<δ⇒∣f(x)−L∣<ε
ε is the allowed distance between f(x) and L, and δ is the allowed distance between x and a. The target ε is given first, and δ answers it. When we find a limit, we look at what happens near a, not at a. That's why we write 0<∣x−a∣, which means we don't use x=a.
Finding δ on a graph
x
y
y=x
2
3
4
4
9
16
5
7
x=2x=4⇒x=4⇒x=16
9−4=5,16−9=7
δ=5
Draw horizontal lines at y=L−ε and y=L+ε to the curve, then vertical lines from the intersections to the x-axis to find the boundary x-values. When the two sides give different distances to a, choose δ to be the distance to the nearer one. For x at x=9 with ε=1, the boundary x-values are 4 and 16, which are 5 and 7 units away from 9. We choose δ=5, because the larger distance 7 would allow inputs like x=2.25, where x=1.5 is more than 1 away from 3. Any smaller positive δ works too.
Showing a number is not the limit
f(x)={1,3,x<0x≥0L=2
ε=21:∣1−2∣∣3−2∣=1≥21=1≥21
To show a number L is not the limit, find one ε>0 where no δ works. For this jump at 0 with L=2, choosing ε=21 means the outputs have to be between 1.5 and 2.5. Every interval around 0 contains inputs where f(x)=1 and inputs where f(x)=3. Both outputs are at distance 1 from 2, which is greater than 21, so no δ>0 works.
Writing an ε-δ proof
∣(2x+1)−5∣=2∣x−2∣<ε⇒δ=2ε
∣(2x+1)−5∣=2∣x−2∣<ε⇒δ=2ε
Let ε>0,δ=2ε:0<∣x−2∣<δ⇒∣(2x+1)−5∣=2∣x−2∣<2δ=ε
Let ε>0,δ=2ε:0<∣x−2∣<δ⇒∣(2x+1)−5∣=2∣x−2∣<2δ=ε
To find δ, work backwards from ∣f(x)−L∣<ε to isolate ∣x−a∣. When ∣f(x)−L∣ simplifies to a constant times ∣x−a∣, divide ε by that constant to get δ. For 2x+1 near 2, factoring out 2 gives 2∣x−2∣<ε, so δ=2ε. Then write the proof forwards. Let ε>0, choose that δ, and show that 0<∣x−a∣<δ leads to ∣f(x)−L∣<ε.
When a factor changes with x
∣x2−9∣=∣x−3∣∣x+3∣
δ≤1⇒5<x+3<7
δ=min(1,7ε)
When ∣f(x)−L∣ factors into ∣x−a∣ times an expression that changes with x, choose a restriction like δ≤1 to keep that factor below a fixed number. For x2 near 3, setting δ≤1 keeps x between 2 and 4, so ∣x+3∣<7. That gives ∣x2−9∣<7∣x−3∣, so choosing δ=min(1,7ε) guarantees ∣x2−9∣<ε.
6
Practice
For Problems 1 and 2, enter your answer as an exact decimal. For Problems 3 to 8, write out your scratch work and proof, then check your steps against the solution.
1
For limx→3(5−2x)=−1 and ε=0.01, find the largest δ such that 0<∣x−3∣<δ implies ∣(5−2x)−(−1)∣<0.01.
How can you factor ∣6−2x∣ into a number times ∣x−3∣?
The slope is −2, so the output changes twice as fast as x. Does δ need to be half of ε or twice ε?
If you got 0.02 Check whether you multiplied by 2 instead of dividing. For x=3.015,∣x−3∣=0.015<0.02, but the output is −1.03, which is 0.03 away from −1.If you got −0.005 Remember that δ is a distance, so it must be positive. Factor out ∣−2∣, which equals 2.
Write the distance from the output to the limit in terms of ∣x−3∣.
∣(5−2x)−(−1)∣=∣6−2x∣=∣−2∣∣x−3∣=2∣x−3∣
We need this distance to be less than 0.01, so divide by 2.
2∣x−3∣<0.01⟺∣x−3∣<0.005⇒δ=0.005
2
The graph of f(x)=x6 near x=3 is shown with the lines y=1.5 and y=2.5. Find the largest δ such that 0<∣x−3∣<δ implies x6−2<0.5.
x
y
2
3
4
5
1.5
2
2.5
y=2.5
y=1.5
y=x6
What x-values give x6=2.5 and x6=1.5?
Which endpoint, x=2.4 or x=4, is nearer to 3?
If you got 1 Check whether δ=1 keeps all outputs within 0.5 of 2. On the left side of 3,x=2.2 is within 1 of 3, but 2.26=1130≈2.73, which is greater than 2.5.
Find the endpointsSet x6 equal to the two bounds on y,2.5 and 1.5.
x6=2.5x6=1.5⇒x=2.4⇒x=4
Choose the nearer distanceFind the distance from 3 to each endpoint, then pick the smaller one so the outputs on both sides stay between 1.5 and 2.5.
3−2.4=0.6,4−3=1⇒δ=0.6
Taking δ=1 would include x=2.2, where the output is above 2.5.
δ=1:x=2.22.26=1130>2.5
3
Use the ε-δ definition to prove that limx→−2(21x+3)=2.
What is the distance ∣x−a∣ when a=−2?
Factor 21 out of 21x+1, then take the absolute value to get a multiple of ∣x+2∣.
If you got ∣x−2∣ in the proof Check the sign of a. Since a=−2, the distance from x to a is ∣x−(−2)∣=∣x+2∣.
If you got δ=2ε While δ=2ε is small enough to work, check how you solved 21∣x+2∣<ε in your scratch work.
Scratch workWrite the distance from x to −2, and simplify the difference between the output and the limit.
∣x−(−2)∣(21x+3)−2=∣x+2∣=21x+1=21∣x+2∣
Choose δWe want 21∣x+2∣<ε, so multiply both sides by 2.
21∣x+2∣<ε⟺∣x+2∣<2ε⇒δ=2ε
Write the proofNow write the steps in the order of the definition.
Let f(x)=x+1 for x<1 and f(x)=5−x for x≥1. Use the ε-δ definition to show that limx→1f(x) is not 4.
x
y
1
2
4
y=x+1
y=5−x
What value does f(x) approach as x approaches 1 from the left?
The left-hand limit is 2 and the proposed limit is 4. Pick ε small enough that 4−ε≥2.
If you got ε=3 Check whether any outputs miss the interval (4−3,4+3)=(1,7). For 0<x<1,f(x)=x+1 is between 1 and 2, which is inside (1,7), so ε=3 is too large to disprove the limit.
Check the one-sided limitsAs x approaches 1 from the left, f(x) approaches 2, not 4. The value f(1)=4 plays no role, because limits require x=1.
x→1−:x+1→2x→1+:5−x→4f(1)=4
Pick an ε for which no δ worksChoose ε=1, which requires 3<f(x)<5. For any δ>0, we can choose an input within that distance to the left of 1, where the output is below 2.
ε=1:3<f(x)<5every δ>0:x=1−2δf(x)=2−2δ<2<3
No positive δ keeps the outputs inside (3,5), so the limit can't be 4.
no δ for ε=1⇒x→1limf(x)=4
7
Use the ε-δ definition to prove that x→2limx1=21.
Find a common denominator for x1−21.
If we choose δ≤1, how small can x get when x is within 1 of 2?
If you got δ=min(1,6ε) When x<3,2x1 is greater than 61, not less than 61. To bound 2∣x∣1 from above, use the lower bound x>1.
Combine over a common denominatorSubtract the fractions, and note that ∣2−x∣=∣x−2∣.
x1−21=2x2−x=2∣x∣∣x−2∣
Bound the denominatorRequire δ≤1 so that x>1. That keeps the denominator from getting too small.
δ≤1:1<x<3⇒2∣x∣>2⇒2∣x∣1<21
Choose δWith 2∣x∣1<21, we need 21δ≤ε, so choose the smaller of 1 and 2ε.
x1−21<21∣x−2∣<21δ≤εwhen δ≤2ε⇒δ=min(1,2ε)
Write the proofCombine x>1 with δ≤2ε to complete the proof.
Let ε>0. Choose δ=min(1,2ε).If 0<∣x−2∣<δ, then x>1 andx1−21=2∣x∣∣x−2∣<21δ≤ε.
8
Use the ε-δ definition to prove that limx→9x=3. (Challenge)
What happens when you multiply x−3 by its conjugate x+3?
What is the smallest possible value of the denominator x+3?
Multiply by the conjugateMultiply and divide by x+3 to turn the numerator into ∣x−9∣.
∣x−3∣=x+3∣x−3∣(x+3)=x+3∣x−9∣
Bound the denominatorBecause x≥0, the denominator is at least 3, so dividing by it gives at most 3∣x−9∣.
x+3≥3⇒∣x−3∣≤3∣x−9∣
Choose δSet δ≤9 so that x>0 and the square root is defined, then solve 3δ≤ε.
δ≤9⇒x>03∣x−9∣<3δ≤εwhen δ≤3ε⇒δ=min(9,3ε)
Write the proofChoose δ=min(9,3ε) and verify that 0<∣x−9∣<δ implies ∣x−3∣<ε.
Let ε>0. Choose δ=min(9,3ε).If 0<∣x−9∣<δ, then x>0 and∣x−3∣≤3∣x−9∣<3δ≤ε.
Notice that for ε=1, this gives δ=3, which is smaller than the largest working value of 5 from the graph. The proof only requires a δ that works, not the biggest one.