Optimal Multiphase Transportation with prescribed momentum (Q4421093)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1967376
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | Optimal Multiphase Transportation with prescribed momentum |
scientific article; zbMATH DE number 1967376 |
Statements
Optimal Multiphase Transportation with prescribed momentum (English)
0 references
19 August 2003
0 references
Monge-Kantorovich problem
0 references
optimal transportation problem
0 references
multiphase flow
0 references
0 references
0 references
0 references
0.7958964109420776
0 references
0.7925366759300232
0 references
0.7832146883010864
0 references
0.7799772620201111
0 references
0.77713942527771
0 references
In this paper results of classical continuum mechanics are given, using a model presented by \textit{J.D. Benamou} and \textit{Y.Brenier} [Numer. Math. 84, No. 3, 375--393 (2000; Zbl 0968.76069)] (in that paper the model was used for numerical purposes). The problem can be stated as follows; given two probability measures \(\rho_0\) and \(\rho_1\) on a set \(D\subset R^d\), find a time and space dependent function \(c(t,x)\) on \([0,1]\times D\) and a \(c\)-square integrable vector field \(v(t,x)\) parallel to \(\partial D\) minimising the cost functional \[ \frac{1}{2}\int | v(t,x)| ^2 c(dt,dx) \] with constraints \[ \partial _t c+\text{ div} (cv)=0 \] and initial boundary conditions \[ c(0,x)=\rho_0(x),\quad c(1,x)=\rho_1(x). \] This problem has been shown in [\textit{Y.Brenier}, C. R. Acad. Sci., Paris, Sér. I 305, 805--808 (1987; Zbl 0652.26017)] to be equivalent to the optimal transportation problem \[ \min_{\mu} \int _{D\times D} \text{ cost} (x,y)d\mu(x,y) \] where the minimum is taken over all probability measures \(\mu\) on \(D\times D\) with projections \(\rho_0\) and \(\rho_1\), equivalence given in the case where the cost function is \(| x-y| ^2/2\). The advantage of the fluid dynamic formulation is the possibility to extend it to a multiphase problem. It is possible to consider a family of phases with given densities \(c(t,x,a)\) labelled by \(a\in A\), and given velocities \(v(t,x,a)\) satisfying the equation \[ \partial _t c+\text{ div} (cv)=0; \] then, with a contraints on the total density given by \[ \rho(t,x)=\sum_{a\in A} c(t,x,a)\leq 1, \] it is possible to prove that there exists a saturation rate in the network that cannot be exceeded. In the present paper, the previous problem is considered with a constraint on the total momentum \[ u(t,x)=\sum_{a\in A} c(t,x,a)v(t,x,a) \] rather than on the total density \(\rho\). It is also proved that if \(D=R^3/Z^3\), a flat \(3\)-dimensional torus, if the phases are equally weighted, the total density is \(1\) and the total momentum is \(0\), then, under suitable initial and final data, the optimal transportation problem always admits a solution.
0 references