Newton’s Method: What Could Go Wrong? Newton’s method works (very) well if |f � | is not too small, |f �� | is not too big, and x0 starts near the solution x. We’re not going to discuss these conditions in detail, but let’s see why they’re there. If f �� is too large the graph would be sharply curved, in which case the tangent line might not be a good approximation to the graph and x1 might not be close to the solution. There are a couple of things that can go wrong if x0 is too far from x1 , which we’ll discuss now. If the error E0 = |x − x0 | is greater than 1 and E1 ∼ E02 , the error of your estimate could actually increase as you apply Newton’s method. In the example f √ (x) = x2 − 5, √ if we had chosen x0 = −2 we would have found the solution − 5 and not 5. This convergence to an unexpected root is illustrated in Fig. 1 y = x2-3 x1 x0 tangent to curve at x = x0 Figure 1: Newton’s method converging to an unexpected root. In the same example, if we chose x0 = 0 then f � (x0 ) = 0 and x1 = x0 − ff�((xx00)) is undeﬁned. Finally, there’s a chance that Newton’s method will cycle back and forth between two value and never converge at all. This failure is illustrated in Fig. 2; x2 = x0 , x3 = x1 , and so forth. Newton’s method is a good way of approximating solutions, but applying it requires some intelligence. You must beware of getting an unexpected result or no result at all. The better your initial guess at the solution, the more likely you are to get a correct result. (x0 , y0 ) (x1 , y1) x0 x1 Figure 2: Newton’s method cycling between x0 and x1 .

Newton’s Method: What Could Go Wrong?

Related Documents

Recommended Documents

Description

Related Documents

Want to Download This Document?

Free up your schedule!

Take 5 seconds to unlock