Random compact set

In mathematics, a random compact set is essentially a compact set-valued random variable. Random compact sets are useful in the study of attractors for random dynamical systems.

Definition

Let (M,d) be a complete separable metric space. Let $\mathcal{K}$ denote the set of all compact subsets of M. The Hausdorff metric h on $\mathcal{K}$ is defined by

$h(K_{1}, K_{2}) := \max \left\{ \sup_{a \in K_{1}} \inf_{b \in K_{2}} d(a, b), \sup_{b \in K_{2}} \inf_{a \in K_{1}} d(a, b) \right\}.$

$(\mathcal{K}, h)$ is also а complete separable metric space. The corresponding open subsets generate a σ-algebra on $\mathcal{K}$, the Borel sigma algebra $\mathcal{B}(\mathcal{K})$ of $\mathcal{K}$.

A random compact set is а measurable function K from а probability space $(\Omega, \mathcal{F}, \mathbb{P})$ into $(\mathcal{K}, \mathcal{B} (\mathcal{K}) )$.

Put another way, a random compact set is a measurable function $K : \Omega \to 2^{M}$ such that K(ω) is almost surely compact and

$\omega \mapsto \inf_{b \in K(\omega)} d(x, b)$

is a measurable function for every $x \in M$.

Discussion

Random compact sets in this sense are also random closed sets as in Matheron (1975). Consequently their distribution is given by the probabilities

$\mathbb{P} (X \cap K = \emptyset)$ for $K \in \mathcal{K}.$

(The distribution of а random compact convex set is also given by the system of all inclusion probabilities $\mathbb{P}(X \subset K).$)

For K = {x}, the probability $\mathbb{P} (x \in X)$ is obtained, which satisfies

$\mathbb{P}(x \in X) = 1 - \mathbb{P}(x \not\in X).$

Thus the covering function p_X is given by

$p_{X} (x) = \mathbb{P} (x \in X)$ for $x \in M.$

Of course, p_X can also be interpreted as the mean of the indicator function $\mathbf{1}_{X}:$

$p_{X} (x) = \mathbb{E} \mathbf{1}_{X} (x).$

The covering function takes values between 0 and 1. The set b_X of all $x \in M$ with p_X(x) > 0 is called the support of X. The set k_X, of all $x \in M$ with p_X(x) = 1 is called the kernel, the set of fixed points, or essential minimum e(X). If $X_1, X_2, \ldots$, is а sequence of i.i.d. random compact sets, then almost surely

$\bigcap_{i=1}^\infty X_i = e(X)$

and $\bigcap_{i=1}^\infty X_i$ converges almost surely to e(X).

References

• Matheron, G. (1975) Random Sets and Integral Geometry. J.Wiley & Sons, New York.
• Molchanov, I. (2005) The Theory of Random Sets. Springer, New York.
• Stoyan D., and H.Stoyan (1994) Fractals, Random Shapes and Point Fields. John Wiley & Sons, Chichester, New York.