m103ex-1-10.mw

Limits and Continuity (Thomas-Finney Section 1.1, page 59, Problem 19

Examine the limiting behavior of  f(x) = x^(1/(1-x))  at x=1..

> f:=x^(1/(1-x));

f := x^(1/(1-x))

We start by calculating the value of f(x) for values of x near 1, to see if we can figure out what f(1) should be.

> for j from 1 to 5 do print(1-(.1)^j, evalf(subs(x=1-(.1)^j,f))) od;

.9, .3486784401

.99, .3660323413

.999, .3676954248

.9999, .3678610464

.99999, .3678776018

> for j from 1 to 5 do print(1+.1^j, evalf(subs(x=1+.1^j,f))) od;

1.1, .3855432894

1.01, .3697112123

1.001, .3680633043

1.0001, .3678978344

1.00001, .3678812806

It appears that as we approach x=1 from either side, f is approaching a value of around .36788 or so.

> plot(f,x=0..2,color=blue,thickness=2);

[Plot]

The graph appears to be continuous, and even though the function is not defined at x=1, it seems as though a limit does exist which would make the function continuous.  We can plot over a smaller domain ("zoom") to get a better idea of the limit:

> plot(f,x=0.8..1.2,color=blue,thickness=2);

[Plot]

It certainly looks like there is a limit. We use Maple to calculate the actual limit:

> limit(f,x=1);

exp(-1)

> evalf(%);

.3678794412

p. 59, Problem 29

Calculatethe average rate of change of f(x) = x^3+1 over the intervals (a) [2,3], and (b) [-1,1]..

Thise is easy in Maple:

> f:=x->x^3+1;

f := proc (x) options operator, arrow; x^3+1 end proc

> aroc1:=(f(3)-f(2))/(3-2);

aroc1 := 19

> aroc2:=(f(1)-f(-1))/(1-(-1));

aroc2 := 1

Section 1.2, page 65, Problem 17

Calculate lim((x-5)/(x^2-25)) as proc (x) options operator, arrow; 5 end proc .  Let's just do this one directly:

> Limit((x-5)/(x^2-25),x=5)=limit((x-5)/(x^2-25),x=5);

Limit((x-5)/(x^2-25), x = 5) = 1/10

Problem 21

Calculate lim((t^2+2-2)/(t^2-1)) as proc (t) options operator, arrow; 1 end proc . Let's explore this one a little more carefully:

> f(t) := (t^2+t-2)/(t^2-1);

f(t) := (t^2+t-2)/(t^2-1)

Let's start by plotting the function and seeing if anything obvious goes wrong.

> plot(f(t), t=0..2,color=blue,thickness=2);

[Plot]

The graph looks continuous, so let's see if we can just plug in t=1.

> subs(t=1,f(t));

Error, numeric exception: division by zero

In fact, note that when we plug in t=1, both the numerator and denominator become zero.  So both polynomials are divisible by (t-1).

> simplify((t^2+t-2)/(t-1));

t+2

> simplify((t^2-1)/(t-1));

t+1

So in fact we can cancel the (t-1) factors and get

> g(t):=simplify((t^2+t-2)/(t^2-1));

g(t) := (t+2)/(t+1)

Now when we try to evaluate g(t) we see that we do indeed get an answer, which will be the limit we seek.

> subs(t=1,g(t));

3/2

Note that we could have just directly solved this problem by doing

> limit(f(t),t=1);

3/2

Section 1.3, page 75, Problem 17

Find an open interval abs(x) < delta in which sqrt(x+1) < Float(1, -1) .

We want to give evidence for the fact that Limit(sqrt(x+1), x = 0) = 1 by finding delta for epsilon = Float(1, -1) .

> f:=x->sqrt(x+1);

f := proc (x) options operator, arrow; sqrt(x+1) end proc

> solve(abs(f(0+delta)-1)<1/10);

RealRange(Open((-19)/100), Open(21/100))

So \we will have abs(f(x)-1) < Float(1, -1) provided abs(delta) < 19/100 .

Section 1.4, page 85,  Problem 15

Compute lim((sqrt(h^2+4*h+5)-sqrt(5))/h) as proc (h) options operator, arrow; 0 end proc .  

Maple makes short work of this:

> Limit((sqrt(h^2+4*h+5)-sqrt(5))/h,h=0)=limit((sqrt(h^2+4*h+5)-sqrt(5))/h,h=0);

Limit(((h^2+4*h+5)^(1/2)-5^(1/2))/h, h = 0) = 2*5^(1/2)/5

Problem 17

Compute lim(x+3)*abs(x+2)/(x+2) as proc (x) options operator, arrow; -2 end proc from the right.

> f:=x->(x+3)*abs(x+2)/(x+2);

f := proc (x) options operator, arrow; (x+3)*abs(x+2)/(x+2) end proc

> limit(f(x),x=-2);

undefined

The limit as itself is undefined, however it is possible for us to compute the left and right handed limits.  

> limit(f(x),x=-2,left);

-1

> limit(f(x),x=1,right);

4

So we can see that the limit failed to exist because the left and right limits were different

Problem 24

Compute lim(1/(x-3)) as proc (x) options operator, arrow; 3 end proc from the right.

> Limit(1/(x-3),x=3,right)=limit(1/(x-3),x=3,right);

Limit(1/(x-3), x = 3, right) = infinity

Section 1.5, page 95, Problem 41

Let g(x) = x for x < -2 while g(x) = b*x^2 for -2 <= x .  For which value of b is this function continuous?

For the function to be continuous at x=-2, the two pieces must agree there:

> solve(subs(x=-2,x)=subs(x=-2,b*x^2),b);

(-1)/2

So the function is continuous for b=-1/2. We can define and plot it to make sure:

> f:=piecewise(x<-2,x,x>=-2,-1/2*x^2);

f := PIECEWISE([x, x < -2], [-x^2/2, -2 <= x])

> plot(f,x=-4..4,color=blue,thickness=2);

[Plot]

Section 1.6, page 102, Problem 7

Find an equation for the  tangent to y = 2*sqrt(x) at the point (1,2).

> f:=x->2*sqrt(x);

f := proc (x) options operator, arrow; 2*sqrt(x) end proc

We need to find the tangent line at (1,2) - so first we get the slope:

> tanslope:=limit((f(1+h)-f(1))/h,h=0);

tanslope := 1

> tanline:=2+tanslope*(x-1);

tanline := 1+x

Now we can plot:

> plot({f(x),tanline},x=-0.1..3,color=blue,thickness=2);

[Plot]

Problem 37  Where does the curve y = x^(1/5) have vertical tangents?

> y:=x^(1/5);

y := x^(1/5)

> plot(y, x=-5..5,color=blue,thickness=2);

[Plot]

It's interesting that we only got the part of the curve for x>0, even though the function exists for x<0 -- to get the whole thing, instead of raising x to the 1/5 power, we'll compute the  5th "surd" (or root) of x:

> y:=surd(x,5);

y := surd(x, 5)

> plot(y,x=-5..5,color=blue,thickness=2);

[Plot]

That's better. It appears that the curve may have a vertical tangent at x=0, but probably does not any other place.  From the definition on p.102, we know that there will be a vertical tangent whenever the following limit is equal to infinity.

> Limit(((x+h)^(1/5)-x^(1/5))/h,h=0)=limit(((x+h)^(1/5)-x^(1/5))/h,h=0);

Limit(((x+h)^(1/5)-x^(1/5))/h, h = 0) = 1/(5*x^(4/5))

This will clearly be infinity if and only if x=0, as we anticipated from the above picture.