Quantitative Observables and Averages in Probabilistic Constraint Programming

  • Authors:
  • Alessandra Di Pierro;Herbert Wiklicky

  • Affiliations:
  • -;-

  • Venue:
  • Selected papers from the Joint ERCIM/Compulog Net Workshop on New Trends in Contraints
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

We investigate notions of observable behaviour of programs which include quantitative aspects of computation along with the most commonly assumed qualitative ones. We model these notions by means of a transition system where transitions occur with a given probability and an associated 'cost' expressing some complexity measure (e.g. running time or, in general, resources consumption). The addition of these quantities allows for a natural formulation of the average behaviour of a program, whose specification and analysis is particularly important in the study of system performance and reliability. It also allows for an average-case analysis of programs' complexity, which can be seen as a semantical counterpart of the average-case asymptotic analysis of algorithms. We base our model on the Concurrent Constraint Programming (CCP) paradigm and we argue that it can be an appropriate base for further developments oriented to the analysis and verification of average properties.