## Statistical Model Checking of Black-Box Probabilistic Systems (2004)

### Cached

### Download Links

- [www-osl.cs.uiuc.edu]
- [osl.cs.uiuc.edu]
- [osl.cs.uiuc.edu]
- [www-osl.cs.uiuc.edu]
- CiteULike
- DBLP

### Other Repositories/Bibliography

Venue: | In 16th conference on Computer Aided Verification (CAV’04), volume 3114 of LNCS |

Citations: | 46 - 7 self |

### BibTeX

@INPROCEEDINGS{Sen04statisticalmodel,

author = {Koushik Sen and Mahesh Viswanathan and Gul Agha},

title = {Statistical Model Checking of Black-Box Probabilistic Systems},

booktitle = {In 16th conference on Computer Aided Verification (CAV’04), volume 3114 of LNCS},

year = {2004},

pages = {202--215},

publisher = {Springer}

}

### Years of Citing Articles

### OpenURL

### Abstract

We propose a new statistical approach to analyzing stochastic systems against specifications given in a sublogic of continuous stochastic logic (CSL). Unlike past numerical and statistical analysis methods, we assume that the system under investigation is an unknown, deployed black-box that can be passively observed to obtain sample traces, but cannot be controlled. Given a set of executions (obtained by Monte Carlo simulation) and a property, our algorithm checks, based on statistical hypothesis testing, whether the sample provides evidence to conclude the satisfaction or violation of a property, and computes a quantitative measure (p-value of the tests) of confidence in its answer; if the sample does not provide statistical evidence to conclude the satisfaction or violation of the property, the algorithm may respond with a "don't know" answer. We implemented our algorithm in a Java-based prototype tool called VeStA, and experimented with the tool using case studies analyzed in [15]. Our empirical results show that our approach may, at least in some cases, be faster than previous analysis methods.

### Citations

639 | A compositional approach to performance modelling
- Hillston
- 1996
(Show Context)
Citation Context ...e traces from any state, making it impossible to generate execution samples on a need basis as is required by the Younes et. al's statistical approach. Despite the success of current analysis methods =-=[10, 13, 12, 15, 8], there is-=- therefore a need to develop methods to analyze stochastic processes that can be applied to deployed, unknown "black-box" systems (systems from which traces cannot be generated from any stat... |

228 | Model checking of probabilistic and nondeterministic systems
- Bianco, Alfaro
- 1995
(Show Context)
Citation Context ...hecked for correctness with respect to the specification using symbolic and numerical methods. Model checkers for di#erent classes of stochastic processes and specification logics have been developed =-=[8, 14, 13, 4, 5, 2, 6]-=-. Although the numerical approach is highly accurate, it su#ers from being computation intensive. An alternate method, proposed by Younes and Simmons [16], is based on Monte Carlo simulation and seque... |

202 | PRISM: Probabilistic Symbolic Model Checker
- Kwiatkowska, Norman, et al.
- 2002
(Show Context)
Citation Context ...hecked for correctness with respect to the specification using symbolic and numerical methods. Model checkers for di#erent classes of stochastic processes and specification logics have been developed =-=[8, 14, 13, 4, 5, 2, 6]-=-. Although the numerical approach is highly accurate, it su#ers from being computation intensive. An alternate method, proposed by Younes and Simmons [16], is based on Monte Carlo simulation and seque... |

201 |
Introduction to Mathematical Statistics
- Hogg, Craig
- 1995
(Show Context)
Citation Context ...e the model checker's answer can be guaranteed to be correct within the required error bounds, we instead compute a quantitative measure of confidence (the p-value, in statistical testing terminology =-=[11]-=-) in the model checker's answer. Our algorithm computes the satisfaction of the desired formula by recursively determining the satisfaction of its subformulas (and the confidence of such an answer) in... |

137 | Approximate symbolic model checking of continuous-time Markov chains
- Baier, Katoen, et al.
- 1999
(Show Context)
Citation Context ...hecked for correctness with respect to the specification using symbolic and numerical methods. Model checkers for di#erent classes of stochastic processes and specification logics have been developed =-=[8, 14, 13, 4, 5, 2, 6]-=-. Although the numerical approach is highly accurate, it su#ers from being computation intensive. An alternate method, proposed by Younes and Simmons [16], is based on Monte Carlo simulation and seque... |

87 | D.: Model-checking for Probabilistic Real-time Systems - Alur, Courcoubetis, et al. |

85 | Symbolic model checking for probabilistic processes
- Baier, Clarke, et al.
- 1997
(Show Context)
Citation Context |

85 | Probabilistic Verification of Discrete Event Systems using Acceptance Sampling - Younes, Simmons - 2002 |

71 |
Model-checking continuous-time markov chains
- Aziz, Sanwal, et al.
(Show Context)
Citation Context ...al results show that our approach may, at least in some cases, be faster than previous analysis methods. 1 Introduction Stochastic models and temporal logics such as continuous stochastic logic (CSL) =-=[1, 3]-=- are widely used to model practical systems and analyze their performance and reliability. There are two primary approaches to analyzing the stochastic behavior of such systems: numerical and statisti... |

58 | Multi Terminal Binary Decision Diagrams to Represent and Analyse
- Hermanns, Meyer-Kayser, et al.
- 1999
(Show Context)
Citation Context ...1 means there is one message at station 1, and s = 1 # a = 0 means that the server is polling station 1. Tandem Queuing Network: This case study is based on a simple tandem queuing network studied in =-=[9]-=-. The model is represented as a CTMC which consists of a M/Cox 2 /1-queue sequentially composed with a M/M/1-queue. We use N to denote the capacity of the queues. The state space is #(N 2 ). We verifi... |

58 | Numerical vs. statistical probabilistic model checking: An empirical study
- Younes, Kwiatkowska, et al.
- 2004
(Show Context)
Citation Context ...property, the algorithm may respond with a "don't know" answer. We implemented our algorithm in a Java-based prototype tool called VeStA, and experimented with the tool using case studies an=-=alyzed in [15]-=-. Our empirical results show that our approach may, at least in some cases, be faster than previous analysis methods. 1 Introduction Stochastic models and temporal logics such as continuous stochastic... |

57 |
Stochastic Petri Net Models of Polling Systems
- Ibe, Trivedi
- 1990
(Show Context)
Citation Context ...e traces from any state, making it impossible to generate execution samples on a need basis as is required by the Younes et. al's statistical approach. Despite the success of current analysis methods =-=[10, 13, 12, 15, 8], there is-=- therefore a need to develop methods to analyze stochastic processes that can be applied to deployed, unknown "black-box" systems (systems from which traces cannot be generated from any stat... |

53 | A Markov Chain Model Checker
- Hermanns, Katoen, et al.
- 2000
(Show Context)
Citation Context |

48 |
Approximate Probabilistic Model Checking
- Hérault, Lassaigne, et al.
- 2004
(Show Context)
Citation Context ...e relevant states are discrete, while we are unlikely to succeed for GSMPs in general. A closely related approach to analyzing stochastic systems based on Monte Carlo simulation is by Herault et. al. =-=[7], which ca-=-n model-check discrete-time Markov chains against properties expressed in an expressively weak logic ("positive LTL"). We have implemented the whole procedure in Java as a prototype tool, ca... |

46 |
Verifying continuoustime Markov chains
- Aziz, Sanwal, et al.
- 1996
(Show Context)
Citation Context ...al results show that our approach may, at least in some cases, be faster than previous analysis methods. 1 Introduction Stochastic models and temporal logics such as continuous stochastic logic (CSL) =-=[1, 3]-=- are widely used to model practical systems and analyze their performance and reliability. There are two primary approaches to analyzing the stochastic behavior of such systems: numerical and statisti... |

36 | Verifying quantitative properties of continuous probabilistic timed automata
- Kwiatkowska, Norman, et al.
- 2000
(Show Context)
Citation Context |

7 | On statistical model checking of probabilistic systems - Sen, Viswanathan, et al. - 2005 |