What Is A Linear Fractional Transformation?
A linear fractional transformation is a function that divides one linear expression by another: $f(x) = \frac{ax+b}{cx+d}$, with the constants tied together by the rule $ad - bc \neq 0$. The name is literal. "Linear" because the top and bottom are each degree-one expressions, and "fractional" because one sits over the other. Its other name, the Möbius transformation, appears whenever the inputs are allowed to be complex numbers.
The four constants $a$, $b$, $c$, $d$ define the whole function. Change them and you get a different transformation; keep the condition $ad - bc \neq 0$ and you always get a genuine, reversible one.
Example 1: Identify the parts of a transformation.
Take $f(x) = \frac{3x+2}{x-1}$. Reading off the constants, $a = 3$, $b = 2$, $c = 1$, $d = -1$.
Now test the condition:
$$ad - bc = (3)(-1) - (2)(1) = -3 - 2 = -5$$
Since $-5 \neq 0$, this is a valid linear fractional transformation. If it helps, notice the top and bottom look exactly like the numerator and denominator you meet when simplifying a rational expression, only here the whole ratio is treated as a single function.
What Is The Domain Of A Linear Fractional Transformation?
The domain is every real number except the one that makes the denominator zero. For $f(x) = \frac{ax+b}{cx+d}$ with $c \neq 0$, that forbidden input is $x = -\frac{d}{c}$.
At $x = -\frac{d}{c}$ the bottom becomes $0$ and the fraction is undefined. Every other real number is allowed.
Example 2: Find the domain of $f(x) = \frac{3x+2}{x-1}$.
Set the denominator to zero and solve:
$$x - 1 = 0 \implies x = 1$$
Final answer: the domain is all real numbers except $x = 1$.
That excluded input is not a flaw in the function. It is where the graph has a vertical asymptote, and it is the reason the real graph of a linear fractional transformation is a hyperbola rather than a straight line. When $c = 0$, there is nothing to exclude, and the function flattens into the ordinary line $f(x) = \frac{a}{d}x + \frac{b}{d}$.
How Do You Write A Linear Fractional Transformation As A Matrix?
Every linear fractional transformation $f(x) = \frac{ax+b}{cx+d}$ corresponds to the $2 \times 2$ matrix built from its four constants:
$$f(x) = \frac{ax+b}{cx+d} \quad \longleftrightarrow \quad \begin{bmatrix} a & b \ c & d \end{bmatrix}$$
The top row holds the numerator's coefficients, and the bottom row holds the denominator's. For $f(x) = \frac{3x+2}{x-1}$, the matrix is $\begin{bmatrix} 3 & 2 \ 1 & -1 \end{bmatrix}$.
This is not a coincidence of notation. The quantity $ad - bc$ from the domain condition is exactly the determinant of that matrix. A valid transformation has a non-zero determinant, which is the same phrase you use for a matrix that can be inverted.
Table: The four numbers of $f(x) = \frac{ax+b}{cx+d}$ and where they sit in the matrix.
Constant | Role in the function | Position in the matrix |
|---|---|---|
$a$ | coefficient of $x$ in the numerator | top-left |
$b$ | constant in the numerator | top-right |
$c$ | coefficient of $x$ in the denominator | bottom-left |
$d$ | constant in the denominator | bottom-right |
How Does Composition Work As Matrix Multiplication?
Applying one linear fractional transformation after another gives another linear fractional transformation, and its matrix is the product of the two matrices. Composing functions becomes multiplying grids.
Example 3: Compose two transformations.
Let $f(x) = \frac{x+1}{x+2}$ and $g(x) = 2x$. Their matrices are $\begin{bmatrix} 1 & 1 \ 1 & 2 \end{bmatrix}$ and $\begin{bmatrix} 2 & 0 \ 0 & 1 \end{bmatrix}$.
First compose the functions directly. Feed $g$ into $f$:
$$f(g(x)) = f(2x) = \frac{2x+1}{2x+2}$$
Now multiply the matrices in the same order:
$$\begin{bmatrix} 1 & 1 \ 1 & 2 \end{bmatrix}\begin{bmatrix} 2 & 0 \ 0 & 1 \end{bmatrix} = \begin{bmatrix} 2 & 1 \ 2 & 2 \end{bmatrix}$$
Read that product back as a function and it says $\frac{2x+1}{2x+2}$, the same answer. The matrix picture turns a messy substitution into one round of matrix multiplication.
How Do You Find The Inverse Of A Linear Fractional Transformation?
The inverse of $f(x) = \frac{ax+b}{cx+d}$ undoes it, and it is found the same way you invert a $2 \times 2$ matrix: swap $a$ and $d$, negate $b$ and $c$. In function form:
$$f^{-1}(x) = \frac{dx - b}{-cx + a}$$
This works precisely because $ad - bc \neq 0$. A zero there would leave the matrix with no inverse, and the function with no way back.
Example 4: Invert $f(x) = \frac{3x+2}{x-1}$.
Here $a = 3$, $b = 2$, $c = 1$, $d = -1$. Substitute into the inverse formula:
$$f^{-1}(x) = \frac{(-1)x - 2}{-(1)x + 3} = \frac{-x-2}{-x+3}$$
Multiply the top and bottom by $-1$ to tidy the signs:
$$f^{-1}(x) = \frac{x+2}{x-3}$$
Final answer: $f^{-1}(x) = \frac{x+2}{x-3}$.
The mechanics match the inverse of a 2×2 matrix step for step, and the idea of a function that reverses another is the same one you meet with ordinary inverse functions.
What Are The Fixed Points Of A Linear Fractional Transformation?
A fixed point is an input the transformation leaves unmoved: a value $x$ where $f(x) = x$. Setting the function equal to its input and clearing the fraction turns the search into a quadratic equation.
Example 5: Find the fixed points of $f(x) = \frac{3x+2}{x-1}$.
Set $f(x) = x$ and multiply both sides by $x - 1$:
$$\frac{3x+2}{x-1} = x \implies 3x + 2 = x(x-1)$$
Expand and move everything to one side:
$$x^2 - 4x - 2 = 0$$
Solve with the quadratic formula:
$$x = \frac{4 \pm \sqrt{16 + 8}}{2} = 2 \pm \sqrt{6}$$
Final answer: the two fixed points are $x = 2 + \sqrt{6}$ and $x = 2 - \sqrt{6}$.
A linear fractional transformation has at most two fixed points (unless it is the identity, which fixes everything). That small number, one or two, is what makes these functions so easy to pin down and so useful as building blocks.
Why Does The ad − bc ≠ 0 Condition Matter?
The condition $ad - bc \neq 0$ is the line between a real transformation and a broken one. It looks like fine print, but it carries the whole meaning of the definition.
A zero value collapses the function. If $ad - bc = 0$, the numerator becomes a fixed multiple of the denominator, so the fraction simplifies to a single constant. A function that returns the same number for every input is not a transformation of anything.
It is the determinant, so it decides invertibility. The quantity $ad - bc$ is the determinant of the matrix $\begin{bmatrix} a & b \ c & d \end{bmatrix}$. A zero determinant means the matrix cannot be inverted, and the transformation cannot be undone.
A non-zero value guarantees one input per output. When $ad - bc \neq 0$, no two inputs ever share an output, so the map is reversible and $f^{-1}$ exists.
Those three points are one idea seen from three sides. Reversible, invertible, and never collapsing to a constant are the same requirement, and $ad - bc \neq 0$ is how you check all three at once.
Who Invented The Linear Fractional Transformation?
These transformations carry the name of a nineteenth-century German mathematician who spent his life studying how shapes and spaces map onto themselves.
Two later mathematicians built these transformations into the heart of modern geometry:
Felix Klein (1849–1925, Germany) made transformations like these the centre of his 1872 Erlangen Program, which reframed geometry itself as the study of what stays the same when you transform a space.
Henri Poincaré (1854–1912, France) used Möbius transformations as the motions of hyperbolic geometry, turning them into a tool for whole new kinds of space.
You can read more about the man behind the name in the MacTutor biography of August Ferdinand Möbius.
Where Is A Linear Fractional Transformation Used In The Real World?
The same four-number rule turns up across science and engineering, usually wherever something needs to be reshaped and then reshaped back.
Signal processing and control engineering: the bilinear transform, a linear fractional transformation, converts continuous-time systems into the discrete-time form that digital filters and controllers actually run on.
Computer graphics and imaging: conformal image warping uses these maps to correct wide-angle and panoramic photos, bending curved lines back to straight ones.
Hyperbolic geometry and relativity: Möbius transformations are the natural motions of the hyperbolic plane, and they describe how the night sky appears to shift for an observer moving near the speed of light.
Number theory: the integer versions form the modular group, the structure underneath continued fractions and modular forms.
Complex analysis: they map circles and lines to circles and lines, which lets mathematicians reshape a hard region into an easy one and solve a problem there instead.
One rule, a ratio of two linear pieces, quietly connects photo software, digital audio, relativity, and pure number theory.
What Are The Most Common Linear Fractional Transformation Mistakes?
Three errors account for most lost marks on this topic, confirmed against the GeeksforGeeks determinant-condition notes, IIT Kanpur lecture notes on Möbius transformations, and the LibreTexts geometry treatment of the same maps.
Forgetting to check $ad - bc \neq 0$.
Where it slips in:
A student writes down a function that looks like a linear fractional transformation and starts computing without testing the condition.
Don't do this:
Do not assume any ratio of two linear expressions qualifies. If $ad - bc = 0$, the function is secretly a constant, and questions about its inverse or fixed points have no meaning.
The correct way:
Compute $ad - bc$ first, every time. Only once it is non-zero should you treat the function as a genuine transformation.
Ignoring the excluded input $x = -\frac{d}{c}$.
Where it slips in:
A student reports the domain as all real numbers, or evaluates the function at the very input that makes the denominator zero.
Don't do this:
Do not treat the transformation as defined everywhere. At $x = -\frac{d}{c}$ the denominator is zero and the output does not exist.
The correct way:
Set the denominator to zero, solve for that single value, and state the domain as every real number except it.
Mishandling the signs when finding the inverse.
Where it slips in:
While inverting, a student negates the wrong entries, or swaps $a$ and $d$ but forgets to negate $b$ and $c$.
Don't do this:
Do not apply the formula from half-memory. A single wrong sign produces a function that fails to undo the original.
The correct way:
Follow the fixed pattern: swap $a$ and $d$, negate $b$ and $c$, giving $f^{-1}(x) = \frac{dx-b}{-cx+a}$. Then check by confirming $f(f^{-1}(x)) = x$ on one test value.
Practice Problems On A Linear Fractional Transformation
Use $f(x) = \frac{2x+1}{x-3}$ unless a problem says otherwise. Answers follow each line.
State the domain of $f(x) = \frac{2x+1}{x-3}$. (
Answer: all real numbers except $x = 3$.)Evaluate $f(1)$.
(Answer: $\frac{2(1)+1}{1-3} = \frac{3}{-2} = -\frac{3}{2}$.)Is $g(x) = \frac{2x+4}{x+2}$ a valid linear fractional transformation?
(Answer: $ad - bc = (2)(2) - (4)(1) = 0$, so no; it collapses to the constant $2$.)Write the matrix form of $f(x) = \frac{2x+1}{x-3}$ and its determinant.
(Answer: $\begin{bmatrix} 2 & 1 \ 1 & -3 \end{bmatrix}$, determinant $= (2)(-3) - (1)(1) = -7$.)Find $f^{-1}(x)$ for $f(x) = \frac{2x+1}{x-3}$.
(Answer: $f^{-1}(x) = \frac{-3x-1}{-x+2} = \frac{3x+1}{x-2}$.)Find the fixed points of $f(x) = \frac{2x+1}{x-3}$.
(Answer: solve $x^2 - 5x - 1 = 0$, giving $x = \frac{5 \pm \sqrt{29}}{2}$.)
Where Should You Go Next After A Linear Fractional Transformation?
This topic sits at the crossroads of functions and matrices, so several natural doors open from here.
Types of functions. See where rational functions like this one fit among the linear, quadratic, and exponential families.
Matrices and determinants. The matrix picture and the determinant $ad - bc$ come straight from here.
Relations and functions. Firm up what it means for a map to be reversible, the property the condition $ad - bc \neq 0$ protects.
If your child is bridging school algebra toward this kind of university-level idea, a live Bhanzu trainer teaches the connection between functions and matrices in the Bhanzu algebra program.
Was this article helpful?
Your feedback helps us write better content