Exploiting interleaving semantics in symbolic state-space generation

DMCA

Exploiting interleaving semantics in symbolic state-space generation

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Gianfranco Ciardo , Andrew S. Miner

Venue:	Formal Methods in System Design
Citations:	16 - 3 self

BibTeX

@INPROCEEDINGS{Ciardo_exploitinginterleaving,
    author = {Gianfranco Ciardo and Andrew S. Miner},
    title = {Exploiting interleaving semantics in symbolic state-space generation},
    booktitle = {Formal Methods in System Design},
    year = {},
    pages = {2007}
}

OpenURL

Abstract

Abstract. Symbolic techniques based on Binary Decision Diagrams (BDDs) are widely employed for reasoning about temporal properties of hardware circuits and synchronous controllers. However, they often perform poorly when dealing with the huge state spaces underlying systems based on interleaving semantics, such as communications protocols and distributed software, which are composed of independently acting subsystems that communicate via shared events. This article shows that the efficiency of state–space exploration techniques using decision diagrams can be drastically improved by exploiting the interleaving semantics underlying many event–based and component–based system models. A new algorithm for symbolically generating state spaces is presented that (i) encodes a model’s state vectors with Multi–valued Decision Diagrams (MDDs) rather than flattening them into BDDs and (ii) partitions the model’s Kronecker–consistent next–state function by event and subsystem, thus enabling multiple lightweight next–state transformations rather than a single heavyweight one. Together, this paves the way for a novel iteration order, called saturation, which replaces the breadth–first search order of traditional algorithms. The resulting saturation algorithm is implemented in the tool SMART, and experimental studies show that it is often several orders of magnitude better in terms of time efficiency, final memory consumption, and peak memory consumption than existing symbolic algorithms.

Keyphrases

symbolic state-space generation decision diagram time efficiency huge state space model kronecker next state function saturation algorithm hardware circuit temporal property state space exploration technique interleaving semantics synchronous controller symbolic algorithm peak memory consumption single heavyweight several order state space binary decision diagram multiple lightweight next state transformation system model model state vector communication protocol many event traditional algorithm experimental study search order new algorithm final memory consumption tool smart symbolic technique novel iteration order

DMCA