Linear Approximations
In this worksheet we'll work with the linear approximation formula 2.33 of our text which says that if f is differentiable near then f(x) is approximately
Tables and graphs will be used to help show where
the approximation is a good one.
This worksheet will use the Maple function option under Enter
Expression of MenuMaple. Before
starting be sure that you have the calculus menu of
MenuMaple open and click in the first command region.
EXAMPLE ONE
Select Enter Expression and click on the function option . The dialogue box will now have three input areas, one for the name of the function, one for the name of the independent variable, and one for the expression. Enter f as the name of the function, x as the independent variable and sqrt(x) as the expression.
> f:=x->sqrt(x);
The output that you will see on the screen is a Maple's mapping notation for functions which will alllow us to use f(x) notation in Maple.
Use the differentiate button to find the derivative of f with respect to x. Call the result df.
> df:=D(f);
We will try to approximate the square root of x for
x near . Recall that to get a linear
approximation to a function f for x near we use
To get a linear approximation for x near 4 use Enter Function to enter a function called Lapprox with x as the independent variable and f(4) + df(4)*(x-4) as the expression.
> Lapprox:=x->f(4)+df(4)*(x-4);
Make a table of values showing x, the actual square root, and the approximate square root found by the linear approximation. Use x between 3.8 and 4.2 in steps of 0.02. To do this use table of values with f(x) and Lapprox(x) as expressions (separated by a comma).
> with(math191):
> valutable([f(x),Lapprox(x)],x=3.8..4.2,'step'=0.02);
x funct1(x) funct2(x) ---------------------------------------------- 3.8 1.949358869 1.950000000 3.820000000 1.954482029 1.955000000 3.840000000 1.959591794 1.960000000 3.860000000 1.964688270 1.965000000 3.880000000 1.969771560 1.970000000 3.900000000 1.974841766 1.975000000 3.920000000 1.979898987 1.980000000 3.940000000 1.984943324 1.985000000 3.960000000 1.989974874 1.990000000 3.980000000 1.994993734 1.995000000 4.000000000 2.000000000 2.000000000 4.020000000 2.004993766 2.005000000 4.040000000 2.009975124 2.010000000 4.060000000 2.014944168 2.015000000 4.080000000 2.019900988 2.020000000 4.100000000 2.024845673 2.025000000 4.120000000 2.029778313 2.030000000 4.140000000 2.034698995 2.035000000 4.160000000 2.039607805 2.040000000 4.180000000 2.044504830 2.045000000 4.200000000 2.049390153 2.050000000
In the table what do you notice f(x) and Lapprox(x) when x gets close to ?
Make a graph of the original f(x) and the linear approximation, Lapprox(x), over the interval from 0 to 6.
> plot([f(x),Lapprox(x)],x=0..6,title=`actual in blue aprroximation in red`,color=[blue,red]);
What happens as x gets closer to xo=4?
EXAMPLE TWO Approximate
the sine of 31 degrees.
Enter g as the name of the function, x as the independent variable,
and sin(x) as the expression.
> g:=x->sin(x);
Now find its derivative with respect to x and give the result the name dg.
> dg:=D(g);
Maple expects angles in radians. Since 31 degrees is near Pi/6, we will find a linear approximation near Pi/6. Enter Lsin as the name of the function, x as the independent variable and
g(Pi/6) + dg(Pi/6)*(x - Pi/6) as the expression.
> Lsin:=x->g(Pi/6)+dg(Pi/6)*(x-Pi/6);
Since one degree is Pi/180 radians, 31 degrees is
Pi/6 + Pi/180. At the prompt below type in ans:=Lsin(Pi/6+Pi/180);
and press enter.
> ans:=Lsin(Pi/6+Pi/180);
The following converts this last result to decimal form.
> ans:=evalf(ans);
Now find the actual value of the sine of 31 degrees,
> Actual:=sin(Pi/6+Pi/180);
and convert it to a decimal.
> Actual:=evalf(Actual);
Now find the absolute value of the difference between Actual and ans.
> error:=abs(Actual - ans);
Now make a graph of g(x) and Lsin(x) for x between 0 and Pi/2. Select scaling on so that the x-axis and y-axis will use the same scale.
> plot([g(x),Lsin(x)],x=0..Pi/2,title=`actual in blue approximate in red`,scaling=constrained,color=[blue,red]);
Dan Symancyk Oct 1995, Oct 1996, June 1997, March 1998, Oct 1998, Feb 1999