Lecture 17 - Piecewise continuous functions and improper integrals

In this lecture, we will review piecewise continuous functions and improper integrals. We will need both of these notions in the next topic: the Laplace Transform. Intuitively, a function is said to be continuous if its graph has no breaks; it can be sketched without without removing the pen from the paper. This intuitive concept is captured by the precise denition below. Denition: (1) A function f (t) is continuous at a point a if: a. f (a) exists, b. limta f (t) exists, c. limta f (t) = f (a). (2) A function is continuous if it is continuous at every a. To see why we need to satisfy all 3 conditions, let us examine the graph of a function f (t) below:

It is intuitively clear that f (t) is NOT continuous at t1 . Which of the three conditions fail? a. f (t1 ) exists ( ) b. limtt1 f (t) does NOT exist since limtt+ f (t) = limtt f (t)! ()
1 1

It is also intuitively clear that f (t) is NOT continuous at t2 . Which of the three conditions fail? a. f (t2 ) exists ( ) b. limtt2 f (t) exists ( ) c. limtt2 f (t) = f (t2 ) () Intuitively, a function is piecewise continuous if it consists of a nite number of continuous pieces with nite discontinuities between them. We will make the idea of nite discontinuity precise below. The function whose graph is given above is not continuous but it is piecewise continuous because it consists of three continuous pieces separated by nite discontinuities. Here is the graph of another piecewise continuous function:

It consists of three continuous pieces separated by nite discontinuities. Note that the function need not be dened at the endpoints of the pieces. Only the limit needs to exist and be nite! Here is another example of a piecewise continuous function: x+4 x<0 x2 0<x<5 f (x) = 7 x5

The function f (x) = 1/x is NOT piecewise continuous:

It consists of two continuous pieces but the discontinuity between them is not nite! The intuitive concept of a piecewise continuous function is captured by the precise denition below.

Denition: A function f (t) is piecewise continuous on an interval [a, b] if the interval can be partitioned by a nite number of points a = t0 < t1 < < tn = b so that a. f (t) is continuous on each subinterval (ti1 , ti ), b. f (t) approaches a nite limit as the endpoints of each subinterval are approached from within the interval. Thus, we call a discontinuity of a function f (t) at a point a nite if both limits lima f (t) and lima+ f (t) are nite. Piecewise continuous functions express many natural relationships that occur in physics, engineering, etc. It is important at this point to remember that the methods we have used so far to solve the dierential equations we encountered all relied in several ways on continuity. For example, the continuity of the functions q(t), r(t), and g(t) was crucial in guaranteeing that there exists a unique solution to a dierential equation of the form y = q(t)y + r(t)y + g(t). The next method we learn will be particularly useful, for example, in solving dierential equations of the form y = q(t)y + r(t)y + g(t) where the coecient functions q(t), r(t), g(t) are piecewise continuous but not continuous. If a function is not continuous, it might not make sense to calculate the area under its graph. One example would be the area under the graph of f (t) = 1/t2 from t = 1 to t = 1. Clearly this area is not nite!

The function f (t) = 1/t2 is not piecewise continuous. A nice feature of piecewise continuous functions is that the notion of area under the graph still makes sense for these. For such functions, we compute the area under the graph simply by adding together the areas under each continuous piece. For example the area under the graph of x x<0 x2 0 < x < 5 f (x) = 7 x5 from x = 6 to x = 8 is:
0 5 8

x + 4 dx +
6 0 6 0 0 6

x2 dx +
5 5 8

7 dx =
5

x + 4 dx +

5 0

x2 dx +

8 5

7 dx = x2 /2 + 4x

+ x3 /3

+ 7x
0

6 + 125/3 + 21 = 15 + 125/3 = 170/3 Next, we review improper integrals. An improper integral is the area under the graph of a function f (x) from x = a onward.

f (x) dx
a

It is computed as the limit of the denite integrals

z a

f (x) dx as z .
z

f (x) dx = lim
a

f (x) dx
a

In many cases, this limit will NOT exist since the area could be innite or nite but oscillating depending on z. If the limit EXISTS, it means the area under the graph of f (x) approaches a xed nite number! z If limz a f (x) dx exists, we say that a f (x) dx converges. Otherwise, we say that a f (x) dx diverges. Let us look at some examples. Problem: Find 2 x2 dx. Solution: Even without any calculation, by just looking at the graph of f (x) = x2 , we should be able to predict that the integral diverges! An obvious prerequisite for the improper integral of a function f (x) to converge is that limx f (x) = 0! Let us do the calculation to conrm this: z 2 z x dx = limz 2 x2 dx = limz x3 /3 = limz (z 3 /3 8/3) = 2 So
2

x2 dx DIVERGES!

Problem: Find 0 et dt. Solution: By looking at the graph of f (t) = et , we see that limt et = 0 and therefore it is at least possible that the integral converges.

t z e dt = limz 0 et dt 0 So 0 et dt CONVERGES! Problem: Find 2 1/x dx.

= limz et

z 0

= limz (ez + e0 ) = limz (ez + 1) = 1

Solution: By looking at the graph of f (x) = 1/x, we see that limx 1/x = 0 and thefore it is at least possible that the integral converges.

1/x dx = limz
2

z 2

1/x dx = limz ln(x)

= limz (ln(z) ln(2)) =

So

1/x dx DIVERGES!

Problem: Find 0 sin(x) dx. Solution: Here we can tell simply by looking at the graph of f (x) = sin(x) that limx sin(x) does not exist.

So

sin(x) dx DIVERGES!

Problem: Does 0 e2t dt converge or diverge? z Solution: We can go ahead and nd limz 0 e2t dt directly as before but we are going to use a dierent method for this problem! We already saw from a previous problem that 0 et dt converges. Now let us compare the graphs of t 2t f (t) = e and g(t) = e :

From the graphs it is clear that the areas under the graph of g(t) from a = 0 onward are smaller than the areas under the graph of f (t) from a = 0 onward. Therefore we conclude that if 0 f (t) dt converges, then so does 0 g(t) dt! Therefore 0 e2t dt CONVERGES! The downside of this method is that it does not give a numerical value for the integral. The method is ONLY good for determining whether the integral converges or diverges. Problem: Does 2 5/x dx converge or diverge? Solution: We will use the same comparative approach as above. We already saw from a previous problem that 2 1/x dx diverges. Now let us compare the graphs of f (x) = 1/x and g(x) = 5/x:

From the graphs it is clear that the area under the graph of g(x) from a = 2 onward is larger than the area under the graph of f (x) from a = 2 onward. Therefore we conclude that if 2 f (x) dx diverges, then so does 2 g(x) dx! Therefore 2 5/x dx DIVERGES! Here is a summary of the Comparison Method we used above. If

a. f (x) 0 and f (x) dx converges, b. g(x) 0 and there is b such that for all x b we have g(x) f (x) (eventually g(x) f (x)), then a g(x) dx converges. If a. f (x) 0 and a f (x) dx diverges, b. there is b such that for all x b we have g(x) f (x) (eventually g(x) f (x)), then a g(x) dx diverges. One of the key points to notice about the method is that the comparison between f (x) and g(x) has to hold after a certain point b and not necessarily starting with a. The reason for this is that the area prior to b is nite anyway and so cannot inuence whether the integral converges or diverges.