Definition:P-Product Metric/General Definition
Definition
Let $M_{1'} = \struct {A_{1'}, d_{1'} }, M_{2'} = \struct {A_{2'}, d_{2'} }, \ldots, M_{n'} = \struct {A_{n'}, d_{n'} }$ be metric spaces.
Let $\ds \AA = \prod_{i \mathop = 1}^n A_{i'}$ be the cartesian product of $A_{1'}, A_{2'}, \ldots, A_{n'}$.
Let $p \in \R_{\ge 1}$.
The $p$-product metric on $\AA$ is defined as:
- $\ds \map {d_p} {x, y} := \paren {\sum_{i \mathop = 1}^n \paren {\map {d_{i'} } {x_i, y_i} }^p}^{\frac 1 p}$
where $x = \tuple {x_1, x_2, \ldots, x_n}, y = \tuple {y_1, y_2, \ldots, y_n} \in \AA$.
The metric space $\MM_p := \struct {\AA, d_p}$ is the $p$-product (space) of $M_{1'}, M_{2'}, \ldots, M_{n'}$.
Real Vector Space
This metric is often found in the context of a real vector space $\R^n$:
The $p$-product metric on $\R^n$ is defined as:
- $\ds \map {d_p} {x, y} := \paren {\sum_{i \mathop = 1}^n \size {x_i - y_i}^p}^{\frac 1 p}$
where $x = \tuple {x_1, x_2, \ldots, x_n}, y = \tuple {y_1, y_2, \ldots, y_n} \in \R^n$.
Special Cases
Some special cases of the $p$-product metric are:
Taxicab Metric
The taxicab metric on $\ds \AA = \prod_{i \mathop = 1}^n A_{i'}$ is defined as:
- $\ds \map {d_1} {x, y} := \sum_{i \mathop = 1}^n \map {d_{i'} } {x_i, y_i}$
where $x = \tuple {x_1, x_2, \ldots, x_n}, y = \tuple {y_1, y_2, \ldots, y_n} \in \AA$.
Euclidean Metric
The Euclidean metric on $\ds \AA = \prod_{i \mathop = 1}^n A_{i'}$ is defined as:
- $\ds \map {d_2} {x, y} := \paren {\sum_{i \mathop = 1}^n \paren {\map {d_{i'} } {x_i, y_i} }^2}^{\frac 1 2}$
where $x = \tuple {x_1, x_2, \ldots, x_n}, y = \tuple {y_1, y_2, \ldots, y_n} \in \AA$.
Chebyshev Distance
The Chebyshev distance on $\ds \AA = \prod_{i \mathop = 1}^n A_i$ is defined as:
- $\ds \map {d_\infty} {x, y} = \max_{i \mathop = 1}^n \set {\map {d_i} {x_i, y_i} }$
where $x = \tuple {x_1, x_2, \ldots, x_n}, y = \tuple {y_1, y_2, \ldots, y_n} \in \AA$.
Standard Discrete Metric
The (standard) discrete metric on $\R^2$ is defined as:
- $\map {d_0} {x, y} := \begin {cases}
0 & : x = y \\ 1 & : \exists i \in \set {1, 2}: x_i \ne y_i \end {cases}$
where $x = \tuple {x_1, x_2}, y = \tuple {y_1, y_2} \in \R^2$.
Also see
- $p$-Product Metric is Metric
- Results about $p$-product metrics can be found here.
Note
Note that while $d_1, d_2, \ldots, d_\infty$ are all topologically equivalent, this is not the case with $d_0$.
Notation
The notation is awkward, because it is necessary to use a indexing subscript for the $n$ metric spaces contributing to the product, and also for the $p$th exponential that defines the metric itself.
Thus the "prime" notation on the $n$ metric spaces.