Atlas home || Conferences | Abstracts | about Atlas

Analysis and Topology, Lviv - 2008
May 26 - June 7, 2008
Ivan Franko National University of Lviv
Lviv, Ukraine

Organizers
M.Zarichnyi, O.Skaskiv, T.Banakh (Lviv University)

View Abstracts
Conference Homepage

Approximation of capacities
by
Inna Hlushak
Precarpathian National University
Coauthors: Oleg Nykyforchyn

Given a capacity c on a metric compactum X we investigate optimal approximations by ∪-capacities (or ∩-capacities) and optimal approximations by capacities on a closed subspace X0 ⊂ X.

On the space MX of upper semicontinuous capacities [3] on a metric compactum (X, d) we consider the following metric :
dM(c, c')= inf
{e > 0 | c(Oe(F))+e ≥ c'(F), c'(Oe(F))+e ≥ c(F) for any closed subset F ⊂ X}
where Oe(F) stands for the closed e-neighborhood of F in X.

We call a capacity c ∈ MX a ∩-capacity (necessity measure) [2] if for all closed subsets A, B of X we have c(A∩B)=min{c(A), c(B)}. Analogously, we call a capacity c ∈ MX a ∪-capacity (also possibility measure [2]), if c(A∪B)=max{c(A), c(B)} for all closed sets A, B ⊂ X.

We denote the set of all ∩-capacities (∪-capacities) on a compactum X by MX ( MX). It is proved [2, 3], that the space MX and its subspaces MX and MX are compacta.

If c is a capacity on X, then the function kX(c) that is determined on the set of all closed subsets X by the formula kX(c)(F)=1-c(X\F) is a capacity as well. It is called the dual or conjugate to c. It is also proved in [2] that a capacity c on a compactum X is a ∩-capacity if and only if the dual capacity kX(c) is a ∪-capacity.

Each capacity c0 on a closed subspace X0 of a space X we extend to X by the formula c(F)=c0(F∩X0) for closed subsets F ⊂ X.

We show that distance from a capacity c to the subspace MX is equal to the infimum of all e ≥ 0 such that the following condition (*) holds:
c(X\OeBae) ≤ a+e for a ∈ I, where Bae={x ∈ X | c(Oe(x)) ≥ a-e}

Here is an algorithm that finds one of ∪-capacities that are nearest to c : we find a least e ≥ 0 such that (*) is valid and use the obtained sets Bae to construct a capacity c0 ∈ MX that is an optimal approximation (in fact it is the greatest of optimal approximations). The capacity c0 is of the form:
c0(F) = sup
{a ∈ I | F∩Bae ≠ ∅}
for any closed subset F of X.

The map kX is an isometry of (MX, dM) onto itself, therefore the capacity c'0 ∈ MX that is nearest to a capacity c ∈ MX can be found as follows : we find the capacity c0 ∈ MX which is nearest to kX(c) and than we go to the dual capacity : c'0=kX(c0).

It is also proved that the distance from a capacity c on X to a subspace MX0 ⊂ MX is equal to e = min{e ≥ 0|c(Oe(X0)) ≥ 1-e, c(X\Oe(X0)) ≤ e}. One of nearest to c ∈ MX capacities on X0 is defined by the formula:
ce = min
(c(Oe(F))+e, 1), F ≠ ∅, 0, F=∅.
A capacity c0 ∈ MX0 is one of nearest to c if and only if
kX(kX(c)e) ≤ c0 ≤ ce.

References

[1] Choquet G. Theory of Capacity, Ann. l'Institute Fourier, 5 (1953-1954), 131-295

[2] Hlushak I.D., Nykyforchyn O.R., Submonads of the capacity monad, 2007, Submitted to Carpathian Journal of Mathematics

[3] Zarichnyi M.M., Nykyforchyn O.R., Functor of capacities in the catagory of compacta, Matem. sb., 2008, 199:2, 3-26 (in Russian)

Date received: April 30, 2008


Copyright © 2008 by the author(s). The author(s) of this document and the organizers of the conference have granted their consent to include this abstract in Atlas Conferences Inc. Document # cawm-35.