Linear codes for sources and source networks: Error exponents, universal coding (Q3663167)

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 3815516
Language Label Description Also known as
default for all languages
No label defined
    English
    Linear codes for sources and source networks: Error exponents, universal coding
    scientific article; zbMATH DE number 3815516

      Statements

      Linear codes for sources and source networks: Error exponents, universal coding (English)
      0 references
      0 references
      1982
      0 references
      independent pairs
      0 references
      Galois field
      0 references
      linear maps
      0 references
      Consider a correlated source consisting of identically distributed, independent pairs \(\{(X_i,Y_i)\}_{i=-\infty}^\infty\) with all the \(X_i\)'s selected from a finite Galois field \(A\) with \(| A|\) elements, and all the \(Y_i\)'s selected from a finite Galois field \(B\) with \(| B|\) elements. For any pair \((R_x,R_y)\) satisfying \(R_x>H(X_0| Y_0)\), \(R_y>H(Y_0| X_0)\), \(R_x+R_y>H(X_0,Y_0)\), the author shows there is \(e(R_x,R_y)>0\) such that for each \(n\) there are integers \(j_n\), \(k_n\), linear maps \(f_n: A^n\to A^{j_n}\) and \(g_n: B^n\to B^{k_n}\), and a map \(h_n:A^{j_n}\times B^{k_n}\to A^n\times B^n\) such that \(n^{-1}j_n\log| A|\to R_x\), \(n^{-1}k_n\log| B|\to R_y\), and \[ \text{Prob}\,[h_n(f_n(X^n),g_n(Y^n))\neq (X^n,Y^n)]\leq \exp [-ne(R_x,R_y)] \] for large \(n\), where \(X^n=(X_1,X_2,\ldots,X^n)\), \(Y^n=(Y_1,Y_2,\ldots,Y^n)\).
      0 references

      Identifiers