families of functions
figure 1: the family of functions with
b=2 and
a=1 (green curve) and 3 (blue curve)
figure 2: the family of functions with
a=3 and
b=1 (green curve) and 3 (blue curve)
figure 3: the family of functions with
a=-1 (green
curve), and -2 (blue curve) and
b=2.
An example of how we might look at families of functions. Let's
consider the family given by
y = a x e-b x
We know that this is a family of functions because of the
parameters a and b (which are unspecified
constants: we know they're constant, but don't know what they
are. And suppose that we're asked to find what functions in the family
have a local minimum at (1,-1).
A good place to start with this type of question is to graph the
function for some values of the parameters. This is done in the
figures to the right. What do you notice? Are we going to get a
local minimum at (1,-1) with these parameter choices?
Nope! We obviously are going to need a to be negative! Good
to know. (Why is this? It's because we know by multiplying a
function by a negative value, which is what taking a to be
negative does, reflects it across the x-axis.)
Ok, how do we find the location of the local minimum to make sure that
it's at (1,-1)? The local minimum will be at a critical
point, which is where y'(x) = 0. So let's
find the derivative and see what that says.
dy/dx = a e-b x
- a b x
e-b x
So if we set this to zero, we get
a e-b x
- a b x
e-b x = 0, so
(a e-b x)
(1 - b x) = 0.
We don't want a = 0, and the exponential can never equal
zero, so this says that x = 1/b. Therefore if we
want the minimum to be at x = 1 we must have
b = 1!
Finally, we want the value at the minimum to be y = -1.
So that says
y(1) = a e-b
= 1, or, with b=1,
a e-1 = -1.
So a = -e !
This says that there is in fact only one member of the family that has
a local minimum at (1,-1), and that member is the function
y = -e x e-x
= -x e-x+1
Check that this works by graphing it on your calculator!
Ok, so what's the point?
- It's often useful to graph a number of functions in a
family of functions to see what they do.
- Sometimes you can see what different parameters do in a family
from those graphs, and then use this information to figure out
what they have to be to get the result that you want. In this
case we got some of that type of information when we established
that a < 0.
- Finally, think about what you want to find. If the problem
doesn't ask for something about slope, maximum or minimum
points, you might not need the function's derivative. If it
does, as it did here, you might. Think about what the problem
is asking for before starting!
scaling, power and exponential functions
Last modified: Fri Mar 11 11:01:58 EST 2005
Comments to:glarose(at)umich(dot)edu
©2005 Gavin LaRose, UM Math Dept.