Compositional Analysis of Expected Delay in Networks of Automata
Abstract
We have succeeded in extending our earlier work on probabilistic I/O automata to obtain a practical method for computing termination probabilities and expected termination times in networks of PIOA. Our method is compositional, avoiding the construction of the network's global state space, and as a result very efficient. This represents the culmination of work we first reported in last year's annual report. Probabilistic I/O Automata (PIOA), a natural extension of the '10 Automaton model of Lynch and Tuttle, have been developed by the Pis to model and analyze distributed systems that exhibit behavior that is statistical or probabilistic in nature and subject to timing constraints. The objectives of this grant are to: (1) continue our investigation of the probabilistic I/O automron model with an emphasis on developing techniques that will permit the probabilistic behavior and delay in distributed systems to be analyzed compositionally; (2) implement the resulting techniques in the Concurrency Factory, a CASE environment for the design, specification, verification, and implementation of asynchronous distributed systems; and 3. perform a technology transfer with Northrop Grumman by applying our techniques to several applications currently of interest to Northrop Grumman.
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 31, 2000
- Accession Number
- ADA388763
Entities
People
- Scott A. Smolka
Organizations
- Stony Brook University