210 likes | 332 Views
Error Control for Probabilistic Model Checking. H åkan L. S. Younes Carnegie Mellon University. Contributions. Framework for expressing correctness guarantees of model-checking algorithms Enables comparison of different algorithms Improves understanding of sampling-based algorithms
E N D
Error Control forProbabilistic Model Checking Håkan L. S. Younes Carnegie Mellon University
Contributions • Framework for expressing correctness guarantees of model-checking algorithms • Enables comparison of different algorithms • Improves understanding of sampling-based algorithms • New sampling-based algorithm for probabilistic model checking • Better error control through undecided results Error Control for Probabilistic Model Checing
Probabilistic Model Checking • Given a model M, a state s, and a property , does hold in s for M? • Model: stochastic discrete event system • Property: probabilistic temporal logic formula arrival departure q “The probability is at least 0.1 that the queuebecomes full within 5 minutes” Error Control for Probabilistic Model Checing
Temporal Stochastic Logic (CSL) • Standard logic operators: , , … • Probabilistic operator: ≥ [] • Holds in state s iff probability is at least for paths satisfying and starting in s • Until: ≤T • Holds over path iff becomes true along within time T, and is true until then Error Control for Probabilistic Model Checing
Property Example • “The probability is at least 0.1 that the queue becomes full within 5 minutes” • ≥0.1[≤5full] Error Control for Probabilistic Model Checing
Possible Results ofModel Checking • Given a state s and a formula , a model-checking algorithm A can: • Accept as true in s (s) • Reject as false in s (s) • Return an undecided result (sI) • An error occurs if: • A rejects when is true (false negative) • A accepts when is false (false positive) Error Control for Probabilistic Model Checing
Ideal Error Control • Bound on false negatives: • Pr[s|s ] • Bound on false positives: • Pr[s|s ] • Bound on undecided results: • Pr[sI] Error Control for Probabilistic Model Checing
Unrealistic Expectations 1 – – s≥ [] s≥ [] Probability of acceptingP≥[] as true in s p Actual probability of holding Error Control for Probabilistic Model Checing
Temporal Stochastic Logic with Indifference Regions (CSL) • Indifference region of width 2 centered around probability thresholds • Probabilistic operator: ≥ [] • Holds in state s if probability is at least + for paths satisfying and starting in s • Does not hold if probability is at most − • “Too close to call” if probability is within distance of Error Control for Probabilistic Model Checing
Error Control forCurrent Solution Methods • Bound on false negatives: • Pr[s|s] • Bound on false positives: • Pr[s|s] • No undecided results: = 0 • Pr[sI] = 0 Error Control for Probabilistic Model Checing
Probabilistic Model Checkingwith Indifference Regions 1 – s≥ [] s≥ [] Probability of acceptingP≥[] as true in s s≥ [] s≥ [] p − + Actual probability of holding Error Control for Probabilistic Model Checing
Hypothesis TestingYounes & Simmons (CAV’02) • Single sampling plan: n, c • Generate n sample execution paths • Accept ≥ [] iff more than c paths satisfy • Probability of accepting ≥ [] as true: • Sequential acceptance sampling Error Control for Probabilistic Model Checing
Statistical EstimationHérault et al. (VMCAI’04) • Estimate p using sample of size n: • Choosing n: • Acceptance condition for ≥ []: Same as single sampling plan n, n + 1! Error Control for Probabilistic Model Checing
Statistical Estimation vs.Hypothesis Testing Error Control for Probabilistic Model Checing
Numerical Transient AnalysisBaier et al. (CAV’00) • Estimate p with truncation error : • Acceptance condition for ≥ []: • Pr[s|s] = 0 • Pr[s|s] = 0 Error Control for Probabilistic Model Checing
Alternative Error Control • Bound on false negatives: • Pr[s|s ] • Bound on false positives: • Pr[s|s ] • Bound on undecided results: • Pr[sI|(s) (s)] Error Control for Probabilistic Model Checing
Probabilistic Model Checkingwith Undecided Results Acceptance probability 1 – Undecided result withprobability at least 1 – – Rejection probability p − + Actual probability of holding Error Control for Probabilistic Model Checing
Statistical Solution Method • Simultaneous acceptance sampling plans • H0: p against H1: p – • H0: p + against H1: p • Combining the results • Accept ≥ [] if H0 and H0 are accepted • Reject ≥ [] if H1 and H1 are accepted • Undecided result otherwise Error Control for Probabilistic Model Checing
20 = 10–2 = 0 15 Verification time (seconds) 10 5 0 14 14.1 14.2 14.3 14.4 14.5 Formula time bound (T) Empirical Evaluation(Symmetric Polling System) serv1 P≥0.5[U ≤Tpoll1] Error Control for Probabilistic Model Checing
Empirical Evaluation(Symmetric Polling System) = = = 10–2 Error Control for Probabilistic Model Checing
Summary • Statistical estimation is never more efficient than hypothesis testing • Statistical methods are randomized algorithms for CSL model checking • Numerical methods are exact algorithms for CSL model checking • New statistical solution method with finer error control ( parameter) Error Control for Probabilistic Model Checing