Template:Short descriptionTemplate:Use American English
In numerical analysis, the secant method is a root-finding algorithm that uses a succession of roots of secant lines to better approximate a root of a function f. The secant method can be thought of as a finite-difference approximation of Newton's method, so it is considered a quasi-Newton method. Historically, it is as an evolution of the method of false position, which predates Newton's method by over 3000 years.<ref>Template:Cite journal</ref>
The methodEdit
The secant method is an iterative numerical method for finding a zero of a function Template:Mvar. Given two initial values Template:Math and Template:Math, the method proceeds according to the recurrence relation
- <math>
x_n
= x_{n-1} - f(x_{n-1}) \frac{x_{n-1} - x_{n-2}}{f(x_{n-1}) - f(x_{n-2})}
= \frac{x_{n-2} f(x_{n-1}) - x_{n-1} f(x_{n-2})}{f(x_{n-1}) - f(x_{n-2})}.
</math>
This is a nonlinear second-order recurrence that is well-defined given Template:Mvar and the two initial values Template:Math and Template:Math. Ideally, the initial values should be chosen close to the desired zero.
Derivation of the methodEdit
Starting with initial values Template:Math and Template:Math, we construct a line through the points Template:Math and Template:Math, as shown in the picture above. In point–point form,<ref>Template:Cite book</ref> the equation of this line is
- <math>y = \frac{f(x_1) - f(x_0)}{x_1 - x_0}(x - x_1) + f(x_1).</math>
The root of this linear function, that is the value of Template:Mvar such that Template:Math is
- <math>x = x_1 - f(x_1) \frac{x_1 - x_0}{f(x_1) - f(x_0)}.</math>
We then use this new value of Template:Mvar as Template:Math and repeat the process, using Template:Math and Template:Math instead of Template:Math and Template:Math. We continue this process, solving for Template:Math, Template:Math, etc., until we reach a sufficiently high level of precision (a sufficiently small difference between Template:Math and Template:Math):
- <math>
\begin{align} x_2 & = x_1 - f(x_1) \frac{x_1 - x_0}{f(x_1) - f(x_0)}, \\[6pt] x_3 & = x_2 - f(x_2) \frac{x_2 - x_1}{f(x_2) - f(x_1)}, \\[6pt] & \,\,\,\vdots \\[6pt] x_n & = x_{n-1} - f(x_{n-1}) \frac{x_{n-1} - x_{n-2}}{f(x_{n-1}) - f(x_{n-2})}. \end{align} </math>
ConvergenceEdit
The iterates <math>x_n</math> of the secant method converge to a root of <math>f</math> if the initial values <math>x_0</math> and <math>x_1</math> are sufficiently close to the root and <math>f</math> is well-behaved. When <math>f</math> is twice continuously differentiable and the root in question is a simple root, i.e., it has multiplicity 1, the order of convergence is the golden ratio <math>\varphi = (1+\sqrt{5})/2 \approx 1.618.</math><ref name=":0">{{#invoke:citation/CS1|citation |CitationClass=web }}</ref> This convergence is superlinear but subquadratic.
If the initial values are not close enough to the root or <math>f</math> is not well-behaved, then there is no guarantee that the secant method converges at all. There is no general definition of "close enough", but the criterion for convergence has to do with how "wiggly" the function is on the interval between the initial values. For example, if <math>f</math> is differentiable on that interval and there is a point where <math>f' = 0</math> on the interval, then the algorithm may not converge.
Comparison with other root-finding methodsEdit
The secant method does not require or guarantee that the root remains bracketed by sequential iterates, like the bisection method does, and hence it does not always converge. The false position method (or {{#invoke:Lang|lang}}) uses the same formula as the secant method. However, it does not apply the formula on <math>x_{n-1}</math> and <math>x_{n-2}</math>, like the secant method, but on <math>x_{n-1}</math> and on the last iterate <math>x_k</math> such that <math>f(x_k)</math> and <math>f(x_{n-1})</math> have a different sign. This means that the false position method always converges; however, only with a linear order of convergence. Bracketing with a super-linear order of convergence as the secant method can be attained with improvements to the false position method (see Regula falsi § Improvements in regula falsi) such as the ITP method or the Illinois method.
The recurrence formula of the secant method can be derived from the formula for Newton's method
- <math>x_n = x_{n-1} - \frac{f(x_{n-1})}{f'(x_{n-1})}</math>
by using the finite-difference approximation, for a small <math>\epsilon =x_{n-1} - x_{n-2} </math>:
<math>f'(x_{n-1}) = \lim_{\epsilon \rightarrow 0} \frac {f(x_{n-1})-f(x_{n-1} - \epsilon)}{\epsilon } \approx \frac{f(x_{n-1}) - f(x_{n-2})}{x_{n-1} - x_{n-2}}</math>
The secant method can be interpreted as a method in which the derivative is replaced by an approximation and is thus a quasi-Newton method.
If we compare Newton's method with the secant method, we see that Newton's method converges faster (order 2 against order the golden ratio φ ≈ 1.6).<ref name=":0" /> However, Newton's method requires the evaluation of both <math>f</math> and its derivative <math>f'</math> at every step, while the secant method only requires the evaluation of <math>f</math>. Therefore, the secant method may sometimes be faster in practice. For instance, if we assume that evaluating <math>f</math> takes as much time as evaluating its derivative and we neglect all other costs, we can do two steps of the secant method (decreasing the logarithm of the error by a factor φ2 ≈ 2.6) for the same cost as one step of Newton's method (decreasing the logarithm of the error by a factor of 2), so the secant method is faster. In higher dimensions, the full set of partial derivatives required for Newton's method, that is, the Jacobian matrix, may become much more expensive to calculate than the function itself. If, however, we consider parallel processing for the evaluation of the derivative or derivatives, Newton's method can be faster in clock time though still costing more computational operations overall.
GeneralizationEdit
Broyden's method is a generalization of the secant method to more than one dimension.
The following graph shows the function f in red and the last secant line in bold blue. In the graph, the x intercept of the secant line seems to be a good approximation of the root of f.
Computational exampleEdit
Below, the secant method is implemented in the Python programming language.
It is then applied to find a root of the function Template:Math with initial points <math>x_0 = 10</math> and <math>x_1 = 30</math>
<syntaxhighlight lang="python3"> def secant_method(f, x0: int, x1: int, iterations: int) -> float:
"""Return the root calculated using the secant method."""
for i in range(iterations):
x2 = x1 - f(x1) * (x1 - x0) / float(f(x1) - f(x0))
x0, x1 = x1, x2
# Apply a stopping criterion here (see below)
return x2
def f_example(x):
return x ** 2 - 612
root = secant_method(f_example, 10, 30, 5)
print(f"Root: {root}") # Root: 24.738633748750722
</syntaxhighlight>
It is very important to have a good stopping criterion above, otherwise, due to limited numerical precision of floating point numbers, the algorithm can return inaccurate results if running for too many iterations. For example, the loop above can stop when one of these is reached first: Template:Mono, or Template:Mono, or Template:Mono. <ref>{{#invoke:citation/CS1|citation |CitationClass=web }}</ref>
NotesEdit
See alsoEdit
ReferencesEdit
External linksEdit
- Secant Method Notes, PPT, Mathcad, Maple, Mathematica, Matlab at Holistic Numerical Methods Institute
- {{#invoke:Template wrapper|{{#if:|list|wrap}}|_template=cite web
|_exclude=urlname, _debug, id |url = https://mathworld.wolfram.com/{{#if:SecantMethod%7CSecantMethod.html}} |title = Secant Method |author = Weisstein, Eric W. |website = MathWorld |access-date = |ref = Template:SfnRef }}