Mo logo [home] [lexicon] [problems] [tests] [courses] [auxiliaries] [notes] [staff] german flag

Mathematics-Online course: Linear Algebra - Matrices - Linear Maps

Linear Maps


[previous page] [next page] [table of contents][page overview]

Let $ V,W$ be $ K$-vector spaces. A map $ \alpha: V \longmapsto W$ is called linear map, if for all $ v_1,v_2,v \in V$ and $ \lambda \in K$ the following holds true:

(i)
$ \alpha(v_1+v_2)=\alpha(v_1)+\alpha(v_2)$
(ii)
$ \alpha(\lambda v)=\lambda \alpha(v)$

As illustrated in the following figure, a rotation is a linear mapping.

\includegraphics[width=.4\linewidth]{b_drehung_bild1} \includegraphics[width=.4\linewidth]{b_drehung_bild2}

The sum $ v = v_1+v_2$ and the vectors $ v_i$ form a triangle the shape of which remains unchanged under rotation. That is, it does not matter whether we form the sum before or after rotation. It is evident that a dilatation by factor $ \lambda$ commutes with any rotation.

In an anologous way we can illustrate that a reflection is also a linear mapping.

\includegraphics[width=.4\linewidth]{b_spiegelung_bild1} \includegraphics[width=.4\linewidth]{b_spiegelung_bild2}

However, a translation of points in the plane is not linear. Neither additivity nor homogeneity are satisfied. For example, for

$\displaystyle \alpha:\ (x_1,x_2)\mapsto (x_1+1,x_2)
$

and

$\displaystyle v_1 = (1,0),\, v_2 = (0,1),\,\lambda=2
$

we have
$\displaystyle \alpha(v_1+v_2) = (2,1)$ $\displaystyle \ne$ $\displaystyle (3,1) = \alpha(v_1) + \alpha(v_2)$  
$\displaystyle \alpha(\lambda v) = (3,0)$ $\displaystyle \ne$ $\displaystyle (4,0) = \lambda \alpha(v)\,
.$  

(Authors: Burkhardt/Höllig/Hörner)

The following table provides some examples of mappings of real-valued functions. For each mapping it is indicated which of the two conditions of linearity is/are satisfied.

Funktion additivity homogeneity
$ f\mapsto f^\prime$ X X
$ f\mapsto \vert f\vert$ - -
$ f\mapsto \int_0^1 f$ X X
$ f\mapsto \max f$ - -
$ f\mapsto f(0)$ X X
$ f\mapsto (\max f+\min f)/2 $ - X

Observe mapping $ \alpha$ which assigns to each complex-valued function the corresponding real part. This mapping is additive but not homogeneous. Here we have $ \alpha(\mathrm{i}f)\neq \mathrm{i}\alpha(f)$.

(Authors: Burkhardt/Höllig/Hörner)

  automatically generated 4/21/2005