A class of inverse problems in transport theory (Q3744674)

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 3979745
Language Label Description Also known as
default for all languages
No label defined
    English
    A class of inverse problems in transport theory
    scientific article; zbMATH DE number 3979745

      Statements

      A class of inverse problems in transport theory (English)
      0 references
      0 references
      1986
      0 references
      neutron transport
      0 references
      eigenfunction expansion
      0 references
      inverse problem
      0 references
      Peierls integral equation
      0 references
      inhomogeneous slab
      0 references
      iterative solution
      0 references
      convergence
      0 references
      The paper deals with the following inverse problem in transport theory: let \(\phi (x)\in L^ 2(-a,a)\), \(\phi\) (x)\(\geq 0\) a.e., be a known solution of the Peierls integral equation in an inhomogeneous slab: \[ (1)\quad \phi (x)=\int^{a}_{-a}c(x')E_ 1(| x-x'|)\phi (x')dx' \] with c(x) unknown. The problem consists in finding c(x) in the sense of best approximation, in other words it consists in determining \(c^*(x)\in L^{\infty}(-a,a)\) such that the solution \(\phi^*(x)\) of eq.(1) with \(c(x)=c^*(x)\) has a minimal distance, in the \(L^ 2\) norm, from the given \(\phi\) (x). Using a convexity argument, the author shows that the optimal solution \(c^*(x)\) exists, under suitable restrictions of \(\phi\) (x), is unique and \(w^*\)-stable, in the sense that \(\phi_ n(x)\to \phi (x)\) in \(L^ 2\) implies \(c^*_ n(x)\to^{w^*}c^*(x)\) (but not \(c^*_ n(x)\to c^*(x)\), in \(L^{\infty}(-a,a)).\) The last part of the paper is devoted to the problem of explicitly determining \(c^*(x)\), assuming that \(\phi^*(x)\) is known. Two methods are given: the first method is based on the expansion of \(\phi^*\) in terms of the eigenfunctions of the integral operator with kernel \(E_ 1(| x-x'|)\) (i.e. the operator in eq.(1) with \(c(x)=1)\). The second method is a more direct one. It is based on a singular integral technique and involves the iterative solution of a related Fredholm integral equation. To ensure convergence, however, an additional condition must be added, namely that the parameter a (the half-thickness of the slab) is not smaller than some constant \(a_ 0>0\).
      0 references

      Identifiers