You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
<p>Newton's method can be used to find the minimum of a function $f(x)$. At a minimum the derivative vanishes, $f'(x^*) = 0$, so minimization reduces to finding the root of $f'(x)$.</p>
<p>Newton's method can be used to find the minimum of a function $f(x)$. At a minimum the derivative vanishes, $f'(x^*) = 0$, so minimization reduces to finding the root of $f'(x)$.</p>
64
60
65
-
<p>At each iterate $x_n$ a quadratic (second-order Taylor) approximation is formed:</p>
<p>The next iterate $x_{n+1}$ is the minimizer of $q(x)$, giving the update rule:</p>
68
-
$$\begin{equation*}
69
-
x_{n+1} = x_{n} - \frac{f'(x_n)}{f''(x_n)}
70
-
\end{equation*}$$
71
-
<p>The green curves show the quadratic approximation at each step. Provided $f''(x_n) > 0$ and $x_0$ is close enough to a local minimum, the method converges rapidly.</p>
72
-
</td></tr>
73
-
</table>
61
+
<p>At each iterate $x_n$ a quadratic (second-order Taylor) approximation is formed:</p>
<p>The next iterate $x_{n+1}$ is the minimizer of $q(x)$, giving the update rule:</p>
64
+
$$\begin{equation*}
65
+
x_{n+1} = x_{n} - \frac{f'(x_n)}{f''(x_n)}
66
+
\end{equation*}$$
67
+
<p>The green curves show the quadratic approximation at each step. Provided $f''(x_n) > 0$ and $x_0$ is close enough to a local minimum, the method converges rapidly.</p>
68
+
</div>
74
69
75
70
</main>
76
71
@@ -289,8 +284,7 @@ <h2>Newton's method for minimization</h2>
0 commit comments