Download PDFOpen PDF in browser

Generating Asymptotically Non-terminant Initial Variable Values for Linear Diagonalizable Programs.

12 pagesPublished: June 19, 2013

Abstract

We present the key notion of "asymptotically non-terminant initial variable values" for non-terminant loop programs. We show that those specific values are directly associated to inital variable values for which the loop program does not terminate.
Considering linear diagonalizable programs, we describe powerful computational methods that generate automatically and symbolically a semi-linear space represented by a linear system of equalities and inequalities. Each element of this space provides us with asymptotically non-terminant initial variable values. Our approach is based on linear algebraic methods and results. We obtain conditions using a decomposition on a specific basis, involving the loop condition and the matrix encoding the instructions of the loop.

Keyphrases: linear algebra, static analysis, termination analysis

In: Laura Kovács and Temur Kutsia (editors). SCSS 2013. 5th International Symposium on Symbolic Computation in Software Science, vol 15, pages 81--92

Links:
BibTeX entry
@inproceedings{SCSS2013:Generating_Asymptotically_Non_terminant_Initial,
  author    = {Rachid Rebiha and Nadir Matringe and Arnaldo Vieira Moura},
  title     = {Generating Asymptotically Non-terminant Initial Variable Values for Linear Diagonalizable Programs.},
  booktitle = {SCSS 2013. 5th International Symposium on Symbolic Computation in Software Science},
  editor    = {Laura Kovacs and Temur Kutsia},
  series    = {EPiC Series in Computing},
  volume    = {15},
  pages     = {81--92},
  year      = {2013},
  publisher = {EasyChair},
  bibsource = {EasyChair, https://easychair.org},
  issn      = {2398-7340},
  url       = {https://easychair.org/publications/paper/Sn18},
  doi       = {10.29007/dq11}}
Download PDFOpen PDF in browser