Download PDFOpen PDF in browser

Decidable linear list constraints

19 pagesPublished: May 4, 2017


We present new results on a constraint satisfaction problem arising from the inference of resource types in automatic amortized analysis for object-oriented programs by Rodriguez and Hofmann.These constraints are essentially linear inequalities between infinite lists of nonnegative rational numbers which are added and compared pointwise. We study the question of satisfiability of a system of such constraints in two variants with significantly different complexity. We show that in its general form (which is the original formulation presented by Hofmann and Rodriguez at LPAR 2012) this satisfiability problem is hard for the famous Skolem-Mahler-Lech problem whose decidability status is still open but which is at least NP-hard. We then identify a subcase of the problem that still covers all instances arising from type inference in the aforementioned amortized analysis and show decidability of satisfiability in polynomial time by a reduction to linear programming. We further give a classification of the growth rates of satisfiable systems in this format and are now able to draw conclusions about resource bounds for programs that involve lists and also arbitrary data structures if we make the additional restriction that their resource annotations are generated by an infinite list (rather than an infinite tree as in the most general case). Decidability of the tree case which was also part of the original formulation by Hofmann and Rodriguez still remains an open problem.

Keyphrases: constraint satisfaction, infinite lists, linear arithmetic, resource types

In: Thomas Eiter and David Sands (editors). LPAR-21. 21st International Conference on Logic for Programming, Artificial Intelligence and Reasoning, vol 46, pages 181--199

BibTeX entry
  author    = {Sabine Bauer and Martin Hofmann},
  title     = {Decidable linear list constraints},
  booktitle = {LPAR-21. 21st International Conference on Logic for Programming, Artificial Intelligence and Reasoning},
  editor    = {Thomas Eiter and David Sands},
  series    = {EPiC Series in Computing},
  volume    = {46},
  pages     = {181--199},
  year      = {2017},
  publisher = {EasyChair},
  bibsource = {EasyChair,},
  issn      = {2398-7340},
  url       = {},
  doi       = {10.29007/d7t4}}
Download PDFOpen PDF in browser