UPPAAL Tutorial. UPPAAL Family

Transcription

1 UPPAAL Tutorial Beyond UPPAAL Alexandre David Paul Pettersson RTSS 05 Classic : real-time verification Cora: real-time scheduling Tron: online real-time testing Tiga: timed game Times: schedulability analysis CoVer: test case generation UPPAAL Family 1

2 Real-time verification Presented today UPPAAL Classic Real Time Scheduling Optimality Reachability Safety UPPAAL Cora Cost Optimal Reachability Analysis 2

3 Real Time Testing Off-line Test Generation On-line Test Generation and Execution UPPAAL Tron Timed Games Optimal winning strategies Controller synthesis UPPAAL Tiga 3

4 Times Schedulability Analysis Schedule synthesis Code synthesis Conformance Testing Test suite generation Coverage observer UPPAAL CoVer 4

5 Open Source Initiatives DBM Library (GPL) Efficient operations on DBMs & federations Subtractions & reduction techniques Ruby binding (with graphical viewer) Used in UPPAAL UTAP (UPPAAL TA parser library, LGPL) Parsing & canonical representation of TA Support for full syntax of UPPAAL TA (xta + xml) Soon GUI XML components TRON UPPAAL Tron Light Controller Example 5

6 Released on May 16, 2004 [Fates 04] UPPAAL Tron Online Testing M Env M Imp TRON Black-box conformance testing of real-time systems. Online generation and execution of timed test traces from given TA model. Explicit modelling of environment allowing for more relevant testing Allowing for more efficient testing (guiding) Env Imp UPPAAL Tron Online Testing TRON 6

7 TRON Online State Estimation Timed Automata Specification State-set explorer: maintain and analyse a set of symbolic states in real time! Z Z 4 Z Z Z 5 Z 3 Z 1 7 Z Z 14 ZZ Z 2 Z 18 Z Z 9 Z 6 Z i! 2.75 O? System Under Test TRON Industrial Application Danfoss Electronic Cooling Controller 7

8 Find a memoryless winning strategy taking controllable edges to reach the Goal that is memoryless Rule: 2-player game, controller can choose only controllable transitions Winning run: reachability states I G safety states I B = Untimed Games B G Controllable Uncontrollable Strategy TIGA Similar with timed constraints Choose controllable transitions with time constraints! Find memoryless winning strategy Algorithm: Timed version of Liu & Smolka 98 Forward reachability + Backward fixed-point computation [CONCUR 05] Timed Games x 1 1 x < 1 2 x:=0 x > x > 1 x 2 B G Controllable Uncontrollable Strategy x 1 TIGA x 1 x 2 x 1 x 1 8

9 Assume Time Optimality Winning Strategy The game is winning We know an upper bound B for the minimal time needed to reach the goal Modification Add a clock t (initially unconstrained) Add the global invariant t B Minimum time required = 2 t Result: TIGA x TIGA Case Study: Production Cell GIVEN System moves S, Controller moves C, and property φ FIND strategy s C such that s C S satisfies φ 9

10 Real-time Scheduling Only Only 1 BroBizz Cheat is is possible (drive close to to car car with with Bizz ) UNSAFE 5 Crossing Times SAFE CAN The THEY Car & MAKE Bridge IT Problem TO SAFE WITHIN 70 MINUTES??? Solve Scheduling Problem using UPPAAL Real-time Scheduling UNSAFE 5 10 SAFE

11 CORA Cost Optimal Scheduling Cost-Rates Fuel Fuel consumed per per time-unit UNSAFE 5 10 SAFE OPTIMAL PLAN HAS ACCUMULATED COST=550 and TOTAL TIME=105! Linearly Priced TA: Optimal Scheduling cost =1 cost =2 x<3 x<3 cost =0 cost+=4 y>2, x<2 a {x:=0} c b CORA Timed Automata + Costs on transitions and locations. Cost of performing transition: Transition cost. Cost of performing delay d: ( d x Location cost). Problem : Find the minimum cost of of reaching location c Trace: ε(2.5) (a,x=y=0) (b,x=y=0) (b,x=y=2.5) x 2 0 Cost of Execution Trace: Sum of costs: = 9 (a,x=0,y=2.5) 11

12 CORA Example: Aircraft Landing cost x <= 5 d+l*(t-t) x >= 4 x=5 e*(t-t) land! cost+=2 x <= 5 x <= 9 cost =3 cost =1 x=5 t E T land! L E earliest landing time T target time L latest time e cost rate for being early l cost rate for being late d fixed cost for being late Planes have to keep separation distance to avoid turbulences caused by preceding planes Runway Example: Aircraft Landing CORA 12

13 How CORA Works CORA Special variables in CORA: cost: the cost as mentioned heur: heuristic value to guide the search rem: lower bound on the remaining time to reach the goal Priced zones [CAV01] Guided search (with the heuristic variable) Branch & bound algorithm to prune the statespace from worse current solutions in practice much fewer states may be explored (compared to non-cost version) 13