Calculus Challenge Problem #7

Solution

Non-Fundamental Functions: When Wrong is Right!

A student was wondering why I took off points for his solution to the following integration
problem.
1

+ 2 x + 53 dx = (12 ) + 2 (1) + 53 ( 02 ) + 2 ( 0 ) + 53 = 3 .

First, note the error in the students work. I patiently showed him that the correct method was
03

x3
1 13
2
2
2
2
5
5
5
5
x
2
x
dx
x
x
1
1
+
+
=
+
+
=
+
+

(
)
(
)
+ ( 0 ) + 3 ( 0) = 3 .

3
3
3
0
3
0 3
3

He, of course, argued that since his way was easier and got the same answer, it was a better
solution. This got me wondering
1

Are there other functions for which

f ( x ) dx = f (1) f ( 0 ) ?

Since the fundamental theorem

of calculus is usually required to evaluate a definite integral, we will call a function for which
b

f ( x ) dx = f ( b ) f ( a ) a non-fundamental function on [a, b].

We see that f ( x ) = x 2 + 2 x + 53

is non-fundamental on [0, 1].

1.

Find a linear function that is non-fundamental on [0, 1].

1

We need to find a function f ( x ) = ax + b so that ax + b dx = ( a + b ) b = a and

0

ax + b dx =

1
2

a + b , so a = 2b . Any function, ( y = 6 x + 3, y = 2 x + , etc) will work. The

basic function is f ( x ) = kx +

2.

k
.
2

Are there any other quadratic functions that are non-fundamental on [0, 1]?

Sure, lots of them. Now we need that ax 2 + bx + c dx = ( a + b + c ) c = a + b and

0

ax

+ bx + c dx = 13 a + 12 b + c , so 13 a + 12 b + c = a + b . So, as long as c = 23 a + 12 b , the quadratic

will be non-fundamental. The simplest solution is to set b = 0, so f ( x ) = kx 2 +

3.

2k
3

Can you find all polynomials that are non-fundamental on [0, 1]?

Generalizing the previous result, we see that there is always a condition on the constant that will
make the sum work. For example, with the 6th degree polynomial
f ( x ) = a0 + a1 x + a2 x 2 + a3 x 3 + a4 x 4 + a5 x 5 + a6 x 6 , we require
a1 2a2 3a3 4a4 5a5 6a6
+
+
+
+
+
. Clearly, we could just as easily choose any other
2
3
4
5
6
7
coefficient instead of the constant. Again, the simplest function satisfying these conditions is
6k
n
f ( x ) = kx 6 +
. In general, the nth degree polynomial f ( x ) = kx n + k
satisfies the
7
n +1
conditions.
a0 =

4.
Must the interval of integration be [0, 1] for polynomials to be non-fundamental? Can
we find a more general result?
No, [0, 1] is convenient, but any other interval will do. Consider f ( x ) = ax n + b on the interval
U

[L, U]. Then ax n + b dx =

L

a
U n +1 Ln +1 ) + b (U L ) must equal a (U n Ln ) . Solving
(
n +1

a
U n +1 Ln +1 ) + b (U L ) = a (U n Ln ) , so
(
n +1
n +1
n +1
n +1
n +1
U n Ln )
U n Ln )
(
(
a (U L )
a (U L )
. For every a, U, and L,
b=a

=a

(U L ) n + 1 (U L )
(U L ) n + 1 (U L )
there is a b that works.
5.
Can functions other than polynomials be non-fundamental on some interval? If so, give
some examples.
The obvious choices are exponentials and sines and cosines. Karl Gross, o Bellville High School
contributed the following example:
e
Consider f ( x ) = ln (1 + x ) + c on [0, 1]. Then 2 ln 2 1 + c = ln 2 ln1 , so c = 1 ln 2 = ln .
2
e
So, f ( x ) = ln (1 + x ) + ln is non-fundamental on [0, 1]. But we know that the Maclaurin
2

expansion for f ( x ) = ln (1 + x ) is ln (1 + x ) = x
a2 =

1
2

, a3 = 13 , .

We have from our earlier work that a0 =

x 2 x3 x 4
xn
+ + " + + " with a0 = 0 , a1 = 1 ,
n
2 3 4

a1 2a2 3a3 4a4 5a5 6a6

makes the function
+
+
+
+
+
2
3
4
5
6
7

non-fundamental, so we must have

n +1
2 1 31
e 1
n ( 1)
ln = (1) + + + " +
+"

3 2 4 3
2 2
n + 1 n

and we can use the non-fundamental property to find the sum of the infinite series.
Reference: Graham, Jeffery A., Self-Integrating Polynomials, The College Mathematics
Journal, September, 2005 as presented in The Calculus Collection edited by Caren
Diefenderfer and Roger Nelson, MAA.