Calc 1 Lecture NotesSection 3.1Page 1 of 9
Section 3.1: Linear Approximations and Newton’s Method
Big Idea #1 (of 2): If you look at a small enough region of a function near a given point, the function can be approximated pretty well by its tangent line at that point.
Here is a graph of the function y = f(x) = x3 –3x2 + 2x + 1 and its tangent line at the point (2,1). The equation of the tangent line is y = 2x – 3.
Here is the same function and tangent line in the range of –1.9x < 2.1. Note how close they are.
In fact, the percent error between the function’s y-value and the tangent line’s y-value at the left endpoint of the interval [1.9, 2.1] is:
,
while the percent error between the function’s y-value and the tangent line’s y-value at the right endpoint of the interval [1.9, 2.1] is
.
Advantages / Uses of linear approximations:
- Many times, errors of the size above are acceptable in exchange for being able to approximate y values of the function with the simple equation of the tangent line. (The closer you are to the point, the better the approximation.)
- Linear approximations are useful if we want to calculate values of a function that a calculator can’t handle with its limited number of digits.
- Linear approximations can be used to interpolate between data points.
Equation for the tangent line (linear approximation) of a function f(x) at x = x0:
y = L(x) = f(x0) + f(x0)(x – x0)
This comes from the point slope form for the tangent line at the point (x0, y0) = (x0, f(x0)) that has slope m = f(x0):
Practice:
- Find a linear approximation to f(x) = sin(x) near x = π/3.
- Find a linear approximation to f(x) = sin(x) near x = 0. Use the linear approximation to obtain an approximate value for x = 0.000 000 000 289 6, and compare to your calculator’s answer.
- How close to 0 do the x values need to be in order for the linear approximation of sin(x) to be accurate to within 10%?
- Find an approximate value for (0.999 999 999 9)½. Compare this to what your calculator returns as an answer.
Note: A good choice for x0 should meet the following two criteria:
- f(x0) and f(x0) should be easy to evaluate exactly
- x0 should be close to the number you are trying to plug into the function
- Use a linear approximation to calculate an approximate value of .
- Use the fact that and to approximate cos(8º) using linear interpolation.
Newton’s Method
Big idea:Newton’s Method is a procedure for finding a numerical approximation for a zero of a function. The method involves finding the tangent line to a function near the zero and then finding where the tangent line crosses the x axis, and then repeating this procedure until a desired precision is achieved. This method is usually much faster than the method of bisections. The general iterative formula for successively better approximations to a zero that results from this method is: .
Example: Let’s try to find the zero near x = 1 of the function f(x) = x4 – 6x2 + 3, whose graph is shown below.
- How do we know there is a zero is in the interval [0,1]?
- Show that the four zeros are given exactly by . Use this expression to write down the value of the zero closest to x = 1 to as many digits as your calculator shows.
- The first step of Newton’s Method is to make a rough guess for the zero. We’ll call our initial guess x0 and choose x0 = 1. Why choose x0 = 1?
- The second step of Newton’s Method is to find the tangent line to the function for the initial guess. Do this now.
- The third step of Newton’s Method is to compute where the tangent line crosses the x axis. Do this now. This crossing point should be a much better approximation of the zero of the function than your initial guess x0. If it is not, then you made a bad initial guess. A picture of the situation through this third step is shown below. It is customary to name this next, better approximation x1. You should find that x1 = 0.75.
Just for comparison, what would the method of bisections give us for a first estimate of the zero after starting from the interval [0, 1]?
The rest of Newton’s Method involves repeating the first three steps, except that instead of making a new guess for the zero each time, you use the zero you got from the tangent line you just found. Also, each time you find a new crossing point for the tangent line, it should get labeled with a sequential subscript. The picture below shows the result of applying Newton’s Method a second time. Here we see x1 = 0.74198717, which is pretty darn close to the actual answer.
Just for comparison, what would the method of bisections give us for a second estimate of the zero?
You might have noticed that it gets to be a pain to write down all those calculator digits…
So, let’s re-do the second step above without relying on the numbers:
This representation should help you see the general formula for Newton’s Method, which is called the Newton-Raphson Formula:
In summary, the steps of Newton’s Method are:
1)Make an initial guess x0 “close” to the real zero.
2)Find the equation of the tangent line for that x0:
L(x) = f(x0) + f ’(x0)(x – x0)
3)Calculate where the tangent line crosses the x axis:
L(x) = f(x0) + f ’(x0)(x – x0) = 0
4)Call this intersection point x1. Go back to step 2 using x1 in place of x0.
5)Repeat to find x2, x3, etc. until you are “close enough.”
To use the Newton-Raphson Formula efficiently, you can program the formula into your graphing calculator and never have to write down a single number until the final answer. See the separate handout with today’s notes, and the image below.
Practice:
- What is the zero of f(x) = x3 + 2?
- Where do the graphs of y = cot x and y = x intersect?
- Approximate .
- Derive an iterative formula for computing the square root of any number.
- Derive an iterative formula for computing the nth root of any number.