Decentralized vs. Monolithic Control of Automata and Weighted Automata

Size: px
Start display at page:

Download "Decentralized vs. Monolithic Control of Automata and Weighted Automata"

Transcription

1 Decentralized vs. Monolithic Control of Automata and Weighted Automata Jan Komenda INSTITUTE of MATHEMATICS Academy of Sciences Czech Republic Institute of Mathematics, Brno branch, November 11, 2016

2 Outline Introduction to Supervisory Control Computational Difficulties Decentralized Control of Monolithical DES Weighted automata Residuation of (max,+) and (min,+) series

3 Supervisory control of automata Motivation for supervisory control Supervisory control as generalization of model checking (verification) Supervisory control solves logical problems of safety and liveness. Control problems in complex distributed plants: manufacturing, computer and communication networks Often encountered: partial observations, modular structure, specification is a property independent from the system Language based and state based framework (equivalent, bidirectional transformations)

4 Automata Generators Definition (Generator) A generator is a structure G = (Q,A,δ,q 0 ) Q is a (finite) set of states A is an alphabet (finite set of events) δ : Q A Q is a partial transition function q 0 is the initial state b a d c

5 Generators Behaviour 1 b a 2 3 e b d 4 5 The language generated by G = (Q,A,δ,q 0 ) is the set L(G) = {w A δ(q 0,w) Q} = {ε,b,bd,bde,a,ab,abe,...}

6 Control Objectives = properties to be imposed by control The main control objectives are: Safety language of the controlled system never exceeds the specification Required behavior language of the controlled system includes at least the minimal required behavior Noblockingness controlled system never deadlocks or livelocks (every reachable state is coreachable from a marked state) (for marked languages)

7 General idea Given a system (plant) G a generator Given a specification K a language or generator The aim is to synthesize a supervisor S such that S/G K supervisor as a map and automaton b c b a d a

8 General idea Given a system (plant) G a generator Given a specification K a language or generator The aim is to synthesize a supervisor S such that S/G K supervisor as a map and automaton b c b a d a

9 General idea Given a system (plant) G a generator Given a specification K a language or generator The aim is to synthesize a supervisor S such that S/G K supervisor as a map and automaton b c b a d a

10 Controlled Generator b a d c A controlled generator is a structure (G,A c,γ), where G is a generator over A, Ac A is the set of controllable events, A u = A \ A c is the set of uncontrollable events, and Γ = {γ A A u γ} is the set of control patterns. Example G = ({1,2,3},{a,b,c,d},δ,1,{1}) A = {a,b,c,d}, A c = {c}, A u = {a,b,d} Γ = {{a,b,d},{a,b,c,d}}

11 Supervisory control b a d c A supervisory control for (G,A c,γ) is a map S : L(G) Γ Example Γ = {{a,b,d},{a,b,c,d}} S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d}

12 Closed-loop system b c b a d a A closed-loop system associated with (G,A c,γ) and S is the minimal language L(S/G) A satisfying ε L(S/G) and if s L(S/G), a S(s), and sa L(G), then sa L(S/G). Example S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d} L(S/G) = {ε,...}

13 Closed-loop system b c b a d a A closed-loop system associated with (G,A c,γ) and S is the minimal language L(S/G) A satisfying ε L(S/G) and if s L(S/G), a S(s), and sa L(G), then sa L(S/G). Example S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d} L(S/G) = {ε,a,...}

14 Closed-loop system b c b a d a A closed-loop system associated with (G,A c,γ) and S is the minimal language L(S/G) A satisfying ε L(S/G) and if s L(S/G), a S(s), and sa L(G), then sa L(S/G). Example S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d} L(S/G) = {ε,a,ab,...}

15 Closed-loop system b c b a d a A closed-loop system associated with (G,A c,γ) and S is the minimal language L(S/G) A satisfying ε L(S/G) and if s L(S/G), a S(s), and sa L(G), then sa L(S/G). Example S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d} L(S/G) = {ε,a,ab,aba,...}

16 Closed-loop system b c b a d a A closed-loop system associated with (G,A c,γ) and S is the minimal language L(S/G) A satisfying ε L(S/G) and if s L(S/G), a S(s), and sa L(G), then sa L(S/G). Example S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d} L(S/G) = {ε,a,ab,aba,abab,...}

17 Closed-loop system b c b a d a A closed-loop system associated with (G,A c,γ) and S is the minimal language L(S/G) A satisfying ε L(S/G) and if s L(S/G), a S(s), and sa L(G), then sa L(S/G). Example S(ε) = S(ab) = S((ab) ) = {a,b,d} S(a) = S(aba) = S(a(ba) ) = {a,b,c,d} for all other strings w, S(w) = {a,b,d} L(S/G) = {ε,a,ab,aba,abab,...}

18 Supervisory control problem Problem Given a controlled generator (G,A c,γ) a specification K L(G) Is there a supervisor S : L(G) Γ such that L(S/G) = K? Example K = {ε,a,ab,aba,abab,...} S((ab) ) = {a,b,d}, S(a(ba) ) = {a,b,c,d}, S(w) = {a,b,d} L(S/G) = K b a c d b a

19 Known results Theorem (Ramadge & Wonham, 1987) Let (G,A c,γ) be a controlled generator, and K L(G) be a specification. There exists a supervisor S : L(G) Γ such that if and only if with respect to L(G). L(S/G) = K specification K is controllable

20 Controllability b a c d b a Definition Given a controlled generator (G,A c,γ) K L(G) Let A u = A \ A c, then K is controllable with respect to L(G) if KA u L(G) K Example {ε,a,ab,aba,...}{a,b,d} L(G) {ε,a,ab,aba,...}

21 Partial Observations Actions (events) partition A = A o A uo disjoint union observable events (A o ) and unobservable (A uo ) events Def. (natural projection) P : A A o : P(ε) = ε, { P(s)a, if a Ao, P(sa) = P(s), if a A uo, for L A : P(L) = w LP(w) Motivation: observations are impossible (e.g. fault) or too costly (sensors required) More generally, mask M : A B for B < A M is surjective and catenative and may have indistinguishable events: M(a 1 ) = M(a 2 ) = b B

22 Observer automaton continued Observer of G : replace events from A uo by ε and determinize (ε- removal). Remarkable similarity with the powerset automaton (over A) used in determinization! Unfortunately number of reachable states denoted Q obs can be exponential in Q. In special cases (P has observer property wrt L(G)): Q obs Q. Theorem. Obs(G) recognize P(L(G)), L(Obs(G)) = P(L(G)).

23 Example (Observer) 1 a τ 2 τ 3 a 4 Figure: An automaton. 1,2 a 2,3,4 a 4 Figure: Its observer.

24 Partial observations: Language properties P : A A o, where A o A... observable events Definition Given a controlled generator (G,A c,γ) K L(G) K is observable with respect to L(G) and P if for s prefix(k) and a A c : sa L(G), s a prefix(k), P(s) = P(s ) sa prefix(k).

25 Outline Introduction to Supervisory Control Computational Difficulties Decentralized Control of Monolithical DES Weighted automata Residuation of (max,+) and (min,+) series

26 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z 1x

27 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z a 2x 1x

28 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z 1x a c 2x 1y

29 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z b 3x 1x a c 2x 1y

30 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z 1x a 3x b c 2x 2y c 1y

31 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z 1x a c 2x b 1y c a 3x 2y

32 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z 3x b c 2x 2y a a c d 1x 1y 1z

33 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z c 3x b c 2x 2y a a c d 1x 1y 1z 3y

34 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z c 3x b b c 2x 2y a a c d 1x 1y 1z 3y

35 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z c 3x 3y b b c d 2x 2y 2z a a c d 1x 1y 1z

36 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z c 3x 3y b b c d 2x 2y 2z a a a c d 1x 1y 1z

37 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z c d 3x 3y 3z b b c d 2x 2y 2z a a a c d 1x 1y 1z

38 How are the large-scale systems constructed? Composed of a number of small subsystems Parallel composition (synchronous product) Example a b c d x y z c d 3x 3y 3z b b b c d 2x 2y 2z a a a c d 1x 1y 1z

39 Size of the composition G = G 1 G 2... G n L(G) = n i=1 P 1 i (L(G i )) Two systems with 3 states result in 3 2 = 9 states n systems with k states = k n 2 n states

40 Size of the composition G = G 1 G 2... G n L(G) = n i=1 P 1 i (L(G i )) Two systems with 3 states result in 3 2 = 9 states n systems with k states = k n 2 n states

41 Size of the composition G = G 1 G 2... G n L(G) = n i=1 P 1 i (L(G i )) Two systems with 3 states result in 3 2 = 9 states n systems with k states = k n 2 n states

42 Size of the composition G = G 1 G 2... G n L(G) = n i=1 P 1 i (L(G i )) Two systems with 3 states result in 3 2 = 9 states n systems with k states = k n 2 n states

43 Coordinator G = G 1 G 2... G n G 1 G 2 Subsystems have something in common shared events = communication Abstract the information coordinator G k The coordinator shares this information among the subsystems G 1 G 2... G n G k

44 Coordinated system main idea G k G 1 G 2... G n

45 Coordinated system main idea G k G 1 G 2... G n G k G 1 G k G 2 G k... G n G k

46 Coordinated system main idea G k G 1 G 2... G n G k G 1 G k G 2 G k... G n G k S 1 /(G 1 G k ) S 2 /(G 2 G k )... S n /(G n G k ) K

47 Coordination control problem Problem Given subsystems G 1 and G 2 over alphabets A 1 and A 2, resp. and a specification K L(G 1 G 2 ) Are there a coordinator G k over A k and supervisors S 1 and S 2 such that L(S 1 /[G 1 (G k )]) L(S 2 /[G 2 (G k )]) = K?

48 Simple example A c = {a 1,a 2,a 3 } p i = prepare a i = access (a unique resource) f i = finish p 1 a p 2 a p 3 a f 1 f 2 f 3 Specification: only one system can access the resource

49 Construction of a coordinator Given subsystems G 1 and G 2 over A 1 and A 2, respectively, and a specification K Construct a coordinator G k over A k as follows: 1. Init A k as the set of shared events 2. Extend A k so that K becomes conditionally decomposable

50 Conditional decomposability Definition A language K (A 1 A 2 ) is conditionally decomposable with respect to alphabets A 1, A 2, A k A 1 A if K = P 1+k (K) P 2+k (K) P 1+k : (A 1 A 2 ) (A 1 A k ) There always exists such a A k Example (Projection) P : {a,b,...,y,z} {e,h,l,o} P(abhdeadiladilasoaf ) = hello

51 Construction of a coordinator Given subsystems G 1 and G 2 over A 1 and A 2, respectively, and a specification K Construct a coordinator G k over A k as follows: 1. Init A k as the set of shared events 2. Extend A k so that K becomes conditionally decomposable 3. Set G k = P k (G 1 ) P k (G 2 )

52 Existential result Theorem There exist supervisors S 1, S 2, S k such that L(S 1 /[G 1 G k ]) L(S 2 /[G 2 G k ]) = K if and only if K is conditionally controllable for generators G 1,G 2,G k Definition: K A is conditionally controllable for G 1,G 2,G k if (i) P 1+k (K) is controllable w.r.t. L(G 1 ) L(G k ) (ii) P 2+k (K) is controllable w.r.t. L(G 2 ) L(G k )

53 LCC and Observer properties Definition (LCC) P : A A 0 is LCC wrt s L if σ u A o A u s.t.p 0 (s)σ u P 0 (L), there is no u (A \ A o ) s.t. suσ u L, or there is u (A u \ A o ) s.t. suσ u L. Definition (Observer) P : A A o, where A o A, isl-observer for L A if, t P(L) and s L, if P(s) t then there is u A s. t. su L and Psu) = t. P(s) t P(L) P P s u L Figure: Natural observer.

54 Supremal sublanguages procedure Theorem Given G 1, G 2, G k, and K conditionally decomposable Define supc 1+k = supc(p 1+k (K),L(G 1 ) L(G k )) supc 2+k = supc(p 2+k (K),L(G 2 ) L(G k )) Let P i+k k be (Pi i+k ) 1 (L i )-observer and LCC for (Pi i+k ) 1 (L i ) and P i+k be LCC for P 1 i+k (L i L k ). Alternatively, let L(G 1 ) L(G k ) and L(G 2 ) L(G k ) are mutually controllable wrt shared uncontrollable events A k,u. Then, supc 1+k supc 2+k = supc(k,l,a u ).

55 Simple example coordinator p 1 a p 2 a p 3 a f 1 f 2 f 3 A k = (A 1 A 2 ) (A 1 A 3 ) (A 2 A 3 ) = /0 Extended A k = {a 1,a 2,a 3 } Set G k = P k (G 1 ) P k (G 2 ) P k (G 3 )

56 Simple example coordinator p 1 a p 2 a p 3 a f 1 f 2 f 3 A k = (A 1 A 2 ) (A 1 A 3 ) (A 2 A 3 ) = /0 Extended A k = {a 1,a 2,a 3 } Set G k = P k (G 1 ) P k (G 2 ) P k (G 3 )

57 Simple example coordinator p 1 a p 2 a p 3 a f 1 f 2 f 3 A k = (A 1 A 2 ) (A 1 A 3 ) (A 2 A 3 ) = /0 Extended A k = {a 1,a 2,a 3 } Set G k = P k (G 1 ) P k (G 2 ) P k (G 3 ) a 1,a 2,a 3 Figure: Coordinator G k it says, communicate if you access the resource 1

58 Simple example supervisors We can compute the supervisors a 2,a 3 a 2,a 3 a 1,a 3 a 1,a 3 a 1,a 2 a 1,a 2 p 1 a p 2 a p 3 a f 1 f 2 f 3 Figure: Supervisors supc 1+k, supc 2+k, and supc 3+k The supervisors communicate via coordinator

59 Simple example supervisors We can compute the supervisors a 2,a 3 a 2,a 3 a 1,a 3 a 1,a 3 a 1,a 2 a 1,a 2 p 1 a p 2 a p 3 a f 1 f 2 f 3 Figure: Supervisors supc 1+k, supc 2+k, and supc 3+k The supervisors communicate via coordinator

60 Simple example supervisors We can compute the supervisors a 2,a 3 a 2,a 3 a 1,a 3 a 1,a 3 a 1,a 2 a 1,a 2 p 1 a p 2 a p 3 a f 1 f 2 f 3 Figure: Supervisors supc 1+k, supc 2+k, and supc 3+k The supervisors communicate via coordinator

61 Multi-level hierarchy Overall coord. group 1 coord. group 2 coord.... group m coord. G 1... G i1 G i G i2 G im G im Group 1 Group 2 Group m

62 Outline Introduction to Supervisory Control Computational Difficulties Decentralized Control of Monolithical DES Weighted automata Residuation of (max,+) and (min,+) series

63 Decentralized control problem Observation and Control tasks are distributed among several supervisors S i, i = 1,...,n Problem (Decentralized control) S i observes A o,i A S i control (can disable) A c,i A specification K L(G) When there exist supervisors S i, i = 1,...,n such that n i=1 L(S i/g i ) = K?

64 Decentralized control of DES Find supervisors S i over alphabets A o,i, i I = {1,...,n} s.t. Remarks L(S 1 S 2 S n G) = K Locals supervisors S i exist if and only if K is coobservable and controllable (K. Rudie, W.M. Wonham (1992). If K is not coobservable then find a controllable and coobservable sublanguage of K Definition.( Coobservability.) K L is coobservable wrt L & (A o,i ) i I if s prefix(k) and a A c, if sa L \ prefix(k), then i I s.t. s with P i (s) = P i (s ), s a prefix(k). Algorithm to check coobservability (decidable in a polynomial time) : Rudie, Willems, (1995).

65 Separability and Decomposability Definition. Separability R is separable wrt (A o,i ) i I if R = i I P i (R) = i I (P i ) 1 P i (R) Lemma. R is separable wrt (A o,i ) i I iff R i A o,i i I, s.t. R = i I R i. Definition. Decomposability A language K is decomposable with respect to alphabets (A i ) n i=1 and L if K = n i=1 P i(k) L.

66 Coobservability does not hold If specification K is not coobservable, computing a coobservable or decomposable sublanguage is a difficult problem We use communications: every K can be made coobservable by enlarging (communicating) local observations This is formalized by the concept of conditional decomposability Note that existence of a decomposable sublanguage is undecidable in general!

67 Separability implies Coobservability Lemma The property A o,i A c A c,i, for i = 1,2,...,n, is equivalent to A o,i A c,j A c,i, for i,j = 1,2,...,n. Proposition. Assume that K is decomposable with respect to (A o,i ) n i=1 and L, and that for i = 1,2,...,n, A o,i A c A c,i. Then K is coobservable with respect to L and (A o,i ) n i=1. Theorem. Assume that A o,i A c A c,i, for i = 1,2,...,n. If K is separable with respect to (A o,i ) n i=1, then K L is coobservable with respect to (A o,i ) n i=1 and L.

68 (over)approximation by modular system Idea: plunge decentralized control problem into coordination control problem by A i = A o,i and A c,i = A o,i A c,i. Note that conditional decomposability is just separability of K with respect to (A o,i A k ) n i=1. Theorem Let A o,i A c A c,i, for i = 1,2,...,n. If K = n i=1 P i(k) (separable) wrt (A o,i ) n i=1, then K L is coobservable wrt (A o,i ) n i=1 and L. Hence, separability implies coobservability

69 1. Constructive results of coordination control Conditional Decomposability A language K is conditionally decomposable with respect to alphabets (A i ) n i=1 and A k, where i j (A i A j ) A k if K = n i=1 P i+k(k), where P i+k : ( n i=1 A i) A i+k, for A i+k = A i A k. Idea If supc i+k are synchronously nonconflicting then n i=1 supc(p i+k(k),l(g i ) L(G k ),A i+k,u ) is controllable with respect to L and A u.

70 Application to decentralized control We over-approximate the plant language by n i=1 P i+k(l). Theorem. Let K = prefixk L = L(G) s.t. K is conditionally decomposable wrt (A i ) n i=1 and A k. If languages in the composition are synchronously nonconflicting then M = n i=1 supc(p i+k(k),p i+k (L),A i+k,u ) K is controllable wrt L and A u, and coobservable wrt L and (A i+k ) n i=1. Remark M is easy to compute (only supremal controllale sublanguages), it is coobsevable by construction.

71 3. Modular control with mutual controllability Definition (Lee, Wong 2002). L i A i, where i = 1,2,...,n, are mutually controllable if for all i,j = 1,2,...,n, L j (A j,u A i ) P j (P i ) 1 (L i ) L j. L P j P i L 1 L 2 Proposition If L i A i, for i = 1,2,...,n, are mutually controllable, then for any K = n i=1 K i L, holds. supc( n i=1 K i, n i=1 L i,a u ) = n i=1 supc(k i,l i,a i,u )

72 Mutual controllability in decentralized control Theorem Let K = prefix(k) L and K = n i=1 P i+k(k). If P i+k (L) and P j+k (L) are mutually controllable, for i,j = 1,2,...,n, then M = n i=1 supc(p i+k(k),p i+k (L),A i+k,u ) is a sublanguage of K controllable wrt L and A u, and coobservable wrt L and (A i+k ) n i=1. Remark Corollary: Optimality If L is CD, i.e. separable wrt (A o,i+k ), i I, then M is coobservable wrt L and (A o,i+k ), i I, and M = SupC(K,L,A uc ) (i.e. is the globally optimal solution)

73 Example K = {aa,ba,bbd,abc}, L = {aac,abc,bac,bbd}, A o,1 = A c,1 = {a,c}, and A o,2 = A c,2 = {b,d}. c 4 3 a 8 6 c c 7 5 b a 1 2 b a 0 b 9 d 10 6 c b a a 1 2 b a 0 b 9 d 10 Generator for the plant Generator for the specification

74 Example continued K is not coobservable wrt L and (A o,i ) 2 i=1 : none of the supervisors can distinguish s = ab and s = ba, and bac K while abc K. Find A k /0 s.t. K is cond. decomposable wrt A 1+k and A 2+k. Take A k = {b} : b is communicated As A 1+k A 2+k = {b}, P 1+k (L) and P 2+k (L) are mutually controllable. supc(p 1+k (K),P 1+k (L),A 1+k,u ) = {aa,ba,bb,ab}, supc(p 2+k (K),P 2+k (L),A 2+k,u ) = {bbd}. Conclusion: K = n i=1 supc(p i+k(k),p i+k (L),A i+k,u ) is coobservable wrt L and {a,b,c} and {b,d}.

75 Remarks Idea: Over-approximate plant in Decentralized Supervisory Control Problem and plunge it into modular control with coordinators Compute a by construction coobservable solution The proposed solution is computationally cheap Communications based on multi-level coordination and extension to conditional (inference) architectures of decentralized control Application to products of timed automata: zone abstractions are not compositional

76 Multilevel Hierarchy Subsystems are organize into groups starting from the lowest level: G k over A k G k1 over A k1 G k2 over A k G km over A km G 1... G i1 G i G i2 G im G im Group I 1 Group I 2 Group I m I j = {i j 1 + 1,i j 1 + 2,...,i j } k l k,l {1,...,m} (A I k A Il ) smaller than k l k,l {1,...,n} A k A l )!

77 Multilevel control motivation Centralized coordination suffers from several problems: For large n too many events must be included in A k! Too many events need to be communicated among all subsystems Coordinator as well as its supervisor are too large Our solution: divide subsystems into groups and associate each group with group coordinators that need much less events Top level coordination then may have much fewer events as well

78 Outline Introduction to Supervisory Control Computational Difficulties Decentralized Control of Monolithical DES Weighted automata Residuation of (max,+) and (min,+) series

79 Motivation for weighted automata Generalize logical automata (outputs are in arbitrary semiring, not just Boolean) Generalize linear systems in semirings (from discrete time to free monoids) Outputs semirings can represent e.g. time (max,+), price (min,+) or probability Infinite state systems: states are vectors over semirings But: decidability issues for even elementary problems

80 Control of (max,+) automata inspired by supervisory control (Max,+) automata: weights in (R { }, max, +). class of Timed Discrete Event (dynamical) Systems (TDES) with synchronization and resource sharing strong expressive power in terms of timed Petri nets: 1-safe TPNs can be viewed as heap models: special (max,+)-automata Supervisory control of (max,+) automata : based on formal power series Tensor products in terms of linear representations Hadamard product in terms of behaviors

81 Algebraic preliminaries: dioids Dioids are idempotent semirings, i.e. an idempotent semigroup (D,, ε), endowed with associative (unit element e), a M : a ε = ε a = ε and distributes over : (i) a,b,c D, (a b) c = (a c) (b c) (ii) a,b,c D, c (a b) = (c a) (c b) Idempotent semigroups have natural order a b iff a b = b. In complete structures : a = n=0 a n, with a 0 = e a D.

82 Dioid of formal power series Dioid (idempotent semiring) of formal power series : R max (A) = {s : A R max }. Notation : s = w A s(w)w R max (A) forms a dioid. For s,s R max (A): and pointwise addition: s s = w A (s(w) s (w))w Cauchy (convolution) multiplication: s s = ( s(u)s (v))w. w A Hadamard product: s s w A (s(w) s (w))w. uv=w

83 Residuation theory Residuation theory generalizes inversion An isotone f : D D, where D is a dioid, is said to be residuated if there exists an isotone map h : D D such that f h Id D and h f Id D. h is unique residual of f, denoted by f. If f is residuated then y D, sup{x D f (x) y} exists and belongs to this subset and is equal to f (y). Example: left and right multiplications are always residuated in complete dioids! Notation. a \y = max{x a x y} and y /a = max{x x a y}.

84 Residuation of Hadamard product Addition and Kleene star are not residuated! But, Hadamard product H y : R max (A) R max (A),s s y is residuated. Moreover, Proposition. H y is residuated and its residual is given by H y(s) = w A (s(w) /y(w))w. Proof. (H y H y)(s) = w A [(s(w) /y(w)) y(w)]w w A s(w)w = s (H y H y )(s) = w A [(s(w) y(w)) /y(w)]w w A s(w)w = s

85 Outline Introduction to Supervisory Control Computational Difficulties Decentralized Control of Monolithical DES Weighted automata Residuation of (max,+) and (min,+) series

86 Basic facts Definition. A rational D-series S is unambiguous if it is recognized by a finite D-automaton G = (α, µ,β) s.t. w supp(s), there exists at most one succesful path in G. Definition. A rational D-series is sequential if it is recognized by a deterministic D-automaton G = (α, µ,β), i.e. G has (i) a single initial state q 0 (α(q) = ε q q 0 ) and (ii) a deterministic transition rel. ( a and q, µ(a)(q,q ) ε for at most one state q ). Theorem. Lombardy and Mairesse (2006) A rational (max,+) series is a rational (min,+) series if and only if it is unambiguous. Consequence. A rational (max,+)series with inverse coefficients is rational (min,+) series if and only if it is unambiguous.

87 Residuation of Hadamard product of series For any S 1,S 2 D(A) and for w A, simply (S 1 /S 2 )(w) = S 1 (w)/s 2 (w). Major problems: (i) S 1 /S 2 need not be (max,+)-rational and (ii) Can we decide whether the controller series S 1 /S 2 is non-negative? max a + b a + b / max a + + b + b a Table: (max,+) product and the corresponding residuation.

88 Motivating example. Example. a/1, b/1 Figure: An automaton recognizing the length of a word. a/1, b/0 a/0, b/1 0 1 Figure: An automaton recognizing the maximum number of occcurences of a letter in a word.

89 Example text. First automaton recognizes S 1 = w A w w. The second S 2 = w A max( w a, w b )w. Note: w A, S 1 (w)/ max S 2 (w) = w max( w a, w b ) = min( w a, w b ). Therefore, S 1 / max S 2 = w A min( w a, w b )w is recognized by automaton (b) seen as a (min,+) automaton. S 1 /S 2 is an ambiguous (min,+) rational series. Hence, this series cannot be a (max,+) rational series.

90 Main theorem Theorem. Let S R min Rat(A) and T R max Rat(A). Then S / max T R min Rat(A); T / min S R max Rat(A). Sketch of the proof. The set of states of A is Q = Q S Q T and we set the following: (p,q) Q S Q T, α (p,q) = (α S ) p /(α T ) q ; a A, (p,q),(r,s) Q S Q T, µ(a) (p,q),(r,s) = µ S (a) (p,r) /µ T (a) (q,s) ; (p,q) Q S Q T, β (p,q) = (β S ) p /(β T ) q.

91 Deciding positivity Proposition. For WA A = (α, µ,β) be a weighted automaton, let S min and S max be the respective (min,+) and (max,+) series recognized by A. One can decide whether S min (w) < 0 (resp. S max (w) > 0) for some w in supp(s min ) (resp. in supp(s max )). Sketch of the proof. S min (w) < 0 for some w supp(s) iff there is a minimal path q 0 a 1 q 1...a n q n s.t. α q0 + n i=1 µ(a i) (qi 1,q i ) + β qn and this sum is negative or there exists a minimal cycle with finite negative weight through some state q j {q 0,...,q n }.

92 Discussion When specification S is (min,+) rational and system T is (max,+) rational, we can decide existence of a non-negative delay-controller, namely the (min,+) rational series S / max T Construction: any (min,+) automaton recognizing S / max T can be transformed to an equivalent (min,+) automaton (α, µ,β) in which all entries of α, µ, and β are either non-negative or equal to However, S / max T that fails to be sequential cannot be used for control Another soluton: determinization

93 Determinization Not all weighted automata can be determinized: no finite state deterministic WA recognizing the same series exist in general For weights in semirings it is worse than in rings Sufficient conditions for determinization exist: twin property (Choffrut), clones property (Kirsten) Sufficient conditions for determinization of series stemming from timed PN: bounded fairness Bounded fairness for automata: ω-languages Determinization can be used to build observers Extension to partial observations and decentralized control

94 Conclusions and Perspectives (Multilevel) coordination control It is based on the top-down decomposition, bottom-up approach also exists Application to decentralized supervisory control Timed systems: abstractions or determinization needed for control

95 Thank You

Decentralized Control of Discrete Event Systems with Multiple Local Specializations 1

Decentralized Control of Discrete Event Systems with Multiple Local Specializations 1 Decentralized Control of Discrete Event Systems with Multiple Local Specializations Shengbing Jiang, Vigyan Chandra, Ratnesh Kumar Department of Electrical Engineering University of Kentucky Lexington,

More information

Supervisory Control: Advanced Theory and Applications

Supervisory Control: Advanced Theory and Applications Supervisory Control: Advanced Theory and Applications Dr Rong Su S1-B1b-59, School of EEE Nanyang Technological University Tel: +65 6790-6042, Email: rsu@ntu.edu.sg EE6226, Discrete Event Systems 1 Introduction

More information

Symbolic Decentralized Supervisory Control

Symbolic Decentralized Supervisory Control Symbolic Decentralized Supervisory Control SYMBOLIC DECENTRALIZED SUPERVISORY CONTROL BY URVASHI AGARWAL, B.Eng. a thesis submitted to the department of computing & software and the school of graduate

More information

Optimal Non-blocking Decentralized Supervisory Control Using G-Control Consistency

Optimal Non-blocking Decentralized Supervisory Control Using G-Control Consistency Optimal Non-blocking Decentralized Supervisory Control Using G-Control Consistency Vahid Saeidi a, Ali A. Afzalian *b, Davood Gharavian c * Phone +982173932626, Fax +982177310425 a,b,c Department of Electrical

More information

On Supervisory Control of Concurrent Discrete-Event Systems

On Supervisory Control of Concurrent Discrete-Event Systems On Supervisory Control of Concurrent Discrete-Event Systems Yosef Willner Michael Heymann March 27, 2002 Abstract When a discrete-event system P consists of several subsystems P 1,..., P n that operate

More information

Bridging the Gap between Reactive Synthesis and Supervisory Control

Bridging the Gap between Reactive Synthesis and Supervisory Control Bridging the Gap between Reactive Synthesis and Supervisory Control Stavros Tripakis University of California, Berkeley Joint work with Ruediger Ehlers (Berkeley, Cornell), Stéphane Lafortune (Michigan)

More information

FORMULAS FOR CALCULATING SUPREMAL CONTROLLABLE AND NORMAL SUBLANGUAGES 1 R. D. Brandt 2,V.Garg 3,R.Kumar 3,F.Lin 2,S.I.Marcus 3, and W. M.

FORMULAS FOR CALCULATING SUPREMAL CONTROLLABLE AND NORMAL SUBLANGUAGES 1 R. D. Brandt 2,V.Garg 3,R.Kumar 3,F.Lin 2,S.I.Marcus 3, and W. M. FORMULAS FOR CALCULATING SUPREMAL CONTROLLABLE AND NORMAL SUBLANGUAGES 1 R. D. Brandt 2,V.Garg 3,R.Kumar 3,F.Lin 2,S.I.Marcus 3, and W. M. Wonham 4 2 Department of ECE, Wayne State University, Detroit,

More information

Language Stability and Stabilizability of Discrete Event Dynamical Systems 1

Language Stability and Stabilizability of Discrete Event Dynamical Systems 1 Language Stability and Stabilizability of Discrete Event Dynamical Systems 1 Ratnesh Kumar Department of Electrical Engineering University of Kentucky Lexington, KY 40506-0046 Vijay Garg Department of

More information

On the Design of Adaptive Supervisors for Discrete Event Systems

On the Design of Adaptive Supervisors for Discrete Event Systems On the Design of Adaptive Supervisors for Discrete Event Systems Vigyan CHANDRA Department of Technology, Eastern Kentucky University Richmond, KY 40475, USA and Siddhartha BHATTACHARYYA Division of Computer

More information

Relative Observability and Coobservability of Timed Discrete-Event Systems

Relative Observability and Coobservability of Timed Discrete-Event Systems TSpace Research Repository tspace.library.utoronto.ca Relative Observability and Coobservability of Timed Discrete-Event Systems Kai Cai, Renyuan Zhang, and W. M. Wonham Version Post-Print/Accepted Manuscript

More information

Hierarchical Interface-Based Decentralized. Supervisory Control

Hierarchical Interface-Based Decentralized. Supervisory Control Hierarchical Interface-Based Decentralized Supervisory Control HIERARCHICAL INTERFACE-BASED DECENTRALIZED SUPERVISORY CONTROL BY HUAILIANG LIU, M.Eng. a thesis submitted to the department of computing

More information

On Properties and State Complexity of Deterministic State-Partition Automata

On Properties and State Complexity of Deterministic State-Partition Automata On Properties and State Complexity of Deterministic State-Partition Automata Galina Jirásková 1, and Tomáš Masopust 2, 1 Mathematical Institute, Slovak Academy of Sciences Grešákova 6, 040 01 Košice, Slovak

More information

Supervisor Localization for Large-Scale Discrete-Event Systems under Partial Observation

Supervisor Localization for Large-Scale Discrete-Event Systems under Partial Observation To appear in the International Journal of Control Vol. 00, No. 00, Month 0XX, 1 1 Supervisor Localization for Large-Scale Discrete-Event Systems under Partial Observation Renyuan Zhang a, Kai Cai b a School

More information

PSPACE-completeness of Modular Supervisory Control Problems

PSPACE-completeness of Modular Supervisory Control Problems PSPACE-completeness of Modular Supervisory Control Problems Kurt Rohloff and Stéphane Lafortune Department of Electrical Engineering and Computer Science The University of Michigan 1301 Beal Ave., Ann

More information

On Controllability and Normality of Discrete Event. Dynamical Systems. Ratnesh Kumar Vijay Garg Steven I. Marcus

On Controllability and Normality of Discrete Event. Dynamical Systems. Ratnesh Kumar Vijay Garg Steven I. Marcus On Controllability and Normality of Discrete Event Dynamical Systems Ratnesh Kumar Vijay Garg Steven I. Marcus Department of Electrical and Computer Engineering, The University of Texas at Austin, Austin,

More information

MODULAR MULTITASKING SUPERVISORY CONTROL OF COMPOSITE DISCRETE-EVENT SYSTEMS. Max H. de Queiroz*, José E. R. Cury**

MODULAR MULTITASKING SUPERVISORY CONTROL OF COMPOSITE DISCRETE-EVENT SYSTEMS. Max H. de Queiroz*, José E. R. Cury** MODULAR MULTITASKING SUPERVISORY CONTROL OF COMPOSITE DISCRETE-EVENT SYSTEMS Max H. de Queiroz*, José E. R. Cury** * GEMM CEFET/SC Florianópolis SC 88020-301 Brazil maxqueiroz@cefetsc.edu.br ** LCMI DAS

More information

Decentralized Control of Discrete Event Systems with Bounded or Unbounded Delay Communication

Decentralized Control of Discrete Event Systems with Bounded or Unbounded Delay Communication Decentralized Control of Discrete Event Systems with Bounded or Unbounded Delay Communication Stavros Tripakis Abstract We introduce problems of decentralized control with communication, where we explicitly

More information

Theory of Computation 4 Non-Deterministic Finite Automata

Theory of Computation 4 Non-Deterministic Finite Automata Theory of Computation 4 Non-Deterministic Finite Automata Frank Stephan Department of Computer Science Department of Mathematics National University of Singapore fstephan@comp.nus.edu.sg Theory of Computation

More information

Decentralized Diagnosis of Discrete Event Systems using Unconditional and Conditional Decisions

Decentralized Diagnosis of Discrete Event Systems using Unconditional and Conditional Decisions Decentralized Diagnosis of Discrete Event Systems using Unconditional and Conditional Decisions Yin Wang, Tae-Sic Yoo, and Stéphane Lafortune Abstract The past decade has witnessed the development of a

More information

Diagnosability Analysis of Discrete Event Systems with Autonomous Components

Diagnosability Analysis of Discrete Event Systems with Autonomous Components Diagnosability Analysis of Discrete Event Systems with Autonomous Components Lina Ye, Philippe Dague To cite this version: Lina Ye, Philippe Dague. Diagnosability Analysis of Discrete Event Systems with

More information

Supervisory control under partial observation is an important problem

Supervisory control under partial observation is an important problem 2576 IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 62, NO. 5, MAY 2017 Technical Notes and Correspondence Supervisor Synthesis for Mealy Automata With Output Functions: A Model Transformation Approach Xiang

More information

Symbolic Hierarchical Interface-based Supervisory Control

Symbolic Hierarchical Interface-based Supervisory Control Symbolic Hierarchical Interface-based Supervisory Control Symbolic Synthesis and Verification of Hierarchical Interface-based Supervisory Control By Raoguang Song, B.Eng. A Thesis Submitted to the School

More information

DECENTRALIZED DIAGNOSIS OF EVENT-DRIVEN SYSTEMS FOR SAFELY REACTING TO FAILURES. Wenbin Qiu and Ratnesh Kumar

DECENTRALIZED DIAGNOSIS OF EVENT-DRIVEN SYSTEMS FOR SAFELY REACTING TO FAILURES. Wenbin Qiu and Ratnesh Kumar DECENTRALIZED DIAGNOSIS OF EVENT-DRIVEN SYSTEMS FOR SAFELY REACTING TO FAILURES Wenbin Qiu and Ratnesh Kumar Department of Electrical and Computer Engineering Iowa State University Ames, IA 50011, U.S.A.

More information

An algebraic characterization of unary two-way transducers

An algebraic characterization of unary two-way transducers An algebraic characterization of unary two-way transducers (Extended Abstract) Christian Choffrut 1 and Bruno Guillon 1 LIAFA, CNRS and Université Paris 7 Denis Diderot, France. Abstract. Two-way transducers

More information

Intersection Based Decentralized Diagnosis: Implementation and Verification

Intersection Based Decentralized Diagnosis: Implementation and Verification Intersection Based Decentralized Diagnosis: Implementation and Verification Maria Panteli and Christoforos N. Hadjicostis Abstract We consider decentralized diagnosis in discrete event systems that are

More information

Definition of Büchi Automata

Definition of Büchi Automata Büchi Automata Definition of Büchi Automata Let Σ = {a,b,...} be a finite alphabet. By Σ ω we denote the set of all infinite words over Σ. A non-deterministic Büchi automaton (NBA) over Σ is a tuple A

More information

REPORT MAS-R0404 DECEMBER

REPORT MAS-R0404 DECEMBER C e n t r u m v o o r W i s k u n d e e n I n f o r m a t i c a MAS Modelling, Analysis and Simulation Modelling, Analysis and Simulation Approximating the minimal-cost sensor-selection for discrete-event

More information

Supervisor Localization of Discrete-Event. Systems with Infinite Behavior

Supervisor Localization of Discrete-Event. Systems with Infinite Behavior Supervisor Localization of Discrete-Event 1 Systems with Infinite Behavior (October 25, 2017) Renyuan Zhang 1, Kai Cai 2 arxiv:1710.08751v1 [cs.sy] 24 Oct 2017 Abstract Recently we developed supervisor

More information

Supervisory Control of Manufacturing Systems with Time Specifications

Supervisory Control of Manufacturing Systems with Time Specifications Supervisory Control of Manufacturing Systems with Time Specifications Alexander Schaub Institute of Automatic Control Engineering Technische Universität München Joint Advanced Student School 2008 Bibliography

More information

Compositions of Tree Series Transformations

Compositions of Tree Series Transformations Compositions of Tree Series Transformations Andreas Maletti a Technische Universität Dresden Fakultät Informatik D 01062 Dresden, Germany maletti@tcs.inf.tu-dresden.de December 03, 2004 1. Motivation 2.

More information

Decentralized Modular Control of Concurrent Fuzzy Discrete Event Systems

Decentralized Modular Control of Concurrent Fuzzy Discrete Event Systems 2010 American Control Conference Marriott Waterfront, Baltimore, MD, USA June 30-July 02, 2010 ThB07.2 Decentralized Modular Control of Concurrent Fuzzy Discrete Event Systems Awantha Jayasiri, George

More information

Semi-asynchronous. Fault Diagnosis of Discrete Event Systems ALEJANDRO WHITE DR. ALI KARIMODDINI OCTOBER

Semi-asynchronous. Fault Diagnosis of Discrete Event Systems ALEJANDRO WHITE DR. ALI KARIMODDINI OCTOBER Semi-asynchronous Fault Diagnosis of Discrete Event Systems ALEJANDRO WHITE DR. ALI KARIMODDINI OCTOBER 2017 NC A&T State University http://www.ncat.edu/ Alejandro White Semi-asynchronous http://techlav.ncat.edu/

More information

Reducing the Supervisory Control of Discrete- Event Systems under Partial Observation

Reducing the Supervisory Control of Discrete- Event Systems under Partial Observation MODARES JOURNAL OF ELECTRICAL ENGINEERING, VOL 16, NO 4, WINTER 2016 29 Reducing the Supervisory Control of Discrete- Event Systems under Partial Observation Vahid Saeidi, Ali A. Afzalian, and Davood Gharavian

More information

3515ICT: Theory of Computation. Regular languages

3515ICT: Theory of Computation. Regular languages 3515ICT: Theory of Computation Regular languages Notation and concepts concerning alphabets, strings and languages, and identification of languages with problems (H, 1.5). Regular expressions (H, 3.1,

More information

A Compositional Approach for Verifying Sampled-Data Supervisory Control

A Compositional Approach for Verifying Sampled-Data Supervisory Control A Compositional Approach for Verifying Sampled-Data Supervisory Control A Compositional Approach for Verifying Sampled-Data Supervisory Control by MAHVASH BALOCH, M.Sc. A Thesis Submitted to the School

More information

Hierarchy among Automata on Linear Orderings

Hierarchy among Automata on Linear Orderings Hierarchy among Automata on Linear Orderings Véronique Bruyère Institut d Informatique Université de Mons-Hainaut Olivier Carton LIAFA Université Paris 7 Abstract In a preceding paper, automata and rational

More information

The State Explosion Problem

The State Explosion Problem The State Explosion Problem Martin Kot August 16, 2003 1 Introduction One from main approaches to checking correctness of a concurrent system are state space methods. They are suitable for automatic analysis

More information

Decentralized Control of Discrete Event Systems with Bounded or Unbounded Delay Communication 1

Decentralized Control of Discrete Event Systems with Bounded or Unbounded Delay Communication 1 Decentralized Control of Discrete Event Systems with Bounded or Unbounded Delay Communication 1 Stavros Tripakis 2 VERIMAG Technical Report TR-2004-26 November 2004 Abstract We introduce problems of decentralized

More information

What Information Really Matters in Supervisor Reduction?

What Information Really Matters in Supervisor Reduction? What Information Really Matters in Supervisor Reduction? Rong Su 1 arxiv:1608.04104v1 [cs.sy] 14 Aug 2016 Abstract To make a supervisor comprehensible to a layman has been a long-lasting goal in the supervisory

More information

Synthesis of Maximally Permissive Non-blocking Supervisors for Partially Observed Discrete Event Systems

Synthesis of Maximally Permissive Non-blocking Supervisors for Partially Observed Discrete Event Systems 53rd IEEE Conference on Decision and Control December 5-7, 24. Los Angeles, California, USA Synthesis of Maximally Permissive Non-blocking Supervisors for Partially Observed Discrete Event Systems Xiang

More information

Advanced Automata Theory 2 Finite Automata

Advanced Automata Theory 2 Finite Automata Advanced Automata Theory 2 Finite Automata Frank Stephan Department of Computer Science Department of Mathematics National University of Singapore fstephan@comp.nus.edu.sg Advanced Automata Theory 2 Finite

More information

Kleene Algebras and Algebraic Path Problems

Kleene Algebras and Algebraic Path Problems Kleene Algebras and Algebraic Path Problems Davis Foote May 8, 015 1 Regular Languages 1.1 Deterministic Finite Automata A deterministic finite automaton (DFA) is a model of computation that can simulate

More information

An Overview of Residuated Kleene Algebras and Lattices Peter Jipsen Chapman University, California. 2. Background: Semirings and Kleene algebras

An Overview of Residuated Kleene Algebras and Lattices Peter Jipsen Chapman University, California. 2. Background: Semirings and Kleene algebras An Overview of Residuated Kleene Algebras and Lattices Peter Jipsen Chapman University, California 1. Residuated Lattices with iteration 2. Background: Semirings and Kleene algebras 3. A Gentzen system

More information

Extremal Solutions of Inequations over Lattices with Applications to Supervisory Control 1

Extremal Solutions of Inequations over Lattices with Applications to Supervisory Control 1 Extremal Solutions of Inequations over Lattices with Applications to Supervisory Control 1 Ratnesh Kumar Department of Electrical Engineering University of Kentucky Lexington, KY 40506-0046 Email: kumar@engr.uky.edu

More information

Representing Arithmetic Constraints with Finite Automata: An Overview

Representing Arithmetic Constraints with Finite Automata: An Overview Representing Arithmetic Constraints with Finite Automata: An Overview Bernard Boigelot Pierre Wolper Université de Liège Motivation Linear numerical constraints are a very common and useful formalism (our

More information

FORMAL LANGUAGES, AUTOMATA AND COMPUTABILITY

FORMAL LANGUAGES, AUTOMATA AND COMPUTABILITY 5-453 FORMAL LANGUAGES, AUTOMATA AND COMPUTABILITY NON-DETERMINISM and REGULAR OPERATIONS THURSDAY JAN 6 UNION THEOREM The union of two regular languages is also a regular language Regular Languages Are

More information

Notes on State Minimization

Notes on State Minimization U.C. Berkeley CS172: Automata, Computability and Complexity Handout 1 Professor Luca Trevisan 2/3/2015 Notes on State Minimization These notes present a technique to prove a lower bound on the number of

More information

Modeling and Stability Analysis of a Communication Network System

Modeling and Stability Analysis of a Communication Network System Modeling and Stability Analysis of a Communication Network System Zvi Retchkiman Königsberg Instituto Politecnico Nacional e-mail: mzvi@cic.ipn.mx Abstract In this work, the modeling and stability problem

More information

Aperiodic languages p. 1/34. Aperiodic languages. Verimag, Grenoble

Aperiodic languages p. 1/34. Aperiodic languages. Verimag, Grenoble Aperiodic languages p. 1/34 Aperiodic languages Dejan Ničković Verimag, Grenoble Aperiodic languages p. 2/34 Table of Contents Introduction Aperiodic Sets Star-Free Regular Sets Schützenberger s theorem

More information

Representation of Supervisory Controls using State Tree Structures, Binary Decision Diagrams, Automata, and Supervisor Reduction

Representation of Supervisory Controls using State Tree Structures, Binary Decision Diagrams, Automata, and Supervisor Reduction Representation of Supervisory Controls using State Tree Structures, Binary Decision Diagrams, Automata, and Supervisor Reduction Wujie Chao 1, Yongmei Gan 2, Zhaoan Wang 3, W. M. Wonham 4 1. School of

More information

We define the multi-step transition function T : S Σ S as follows. 1. For any s S, T (s,λ) = s. 2. For any s S, x Σ and a Σ,

We define the multi-step transition function T : S Σ S as follows. 1. For any s S, T (s,λ) = s. 2. For any s S, x Σ and a Σ, Distinguishability Recall A deterministic finite automaton is a five-tuple M = (S,Σ,T,s 0,F) where S is a finite set of states, Σ is an alphabet the input alphabet, T : S Σ S is the transition function,

More information

CS 154. Finite Automata, Nondeterminism, Regular Expressions

CS 154. Finite Automata, Nondeterminism, Regular Expressions CS 54 Finite Automata, Nondeterminism, Regular Expressions Read string left to right The DFA accepts a string if the process ends in a double circle A DFA is a 5-tuple M = (Q, Σ, δ, q, F) Q is the set

More information

Masked Prioritized Synchronization for Interaction and Control of Discrete Event Systems

Masked Prioritized Synchronization for Interaction and Control of Discrete Event Systems Masked Prioritized Synchronization for Interaction and Control of Discrete Event Systems Ratnesh Kumar Department of Electrical Engineering University of Kentucky Lexington, KY 40506-0046 Michael Heymann

More information

Lecture Notes On THEORY OF COMPUTATION MODULE -1 UNIT - 2

Lecture Notes On THEORY OF COMPUTATION MODULE -1 UNIT - 2 BIJU PATNAIK UNIVERSITY OF TECHNOLOGY, ODISHA Lecture Notes On THEORY OF COMPUTATION MODULE -1 UNIT - 2 Prepared by, Dr. Subhendu Kumar Rath, BPUT, Odisha. UNIT 2 Structure NON-DETERMINISTIC FINITE AUTOMATA

More information

CPSC 421: Tutorial #1

CPSC 421: Tutorial #1 CPSC 421: Tutorial #1 October 14, 2016 Set Theory. 1. Let A be an arbitrary set, and let B = {x A : x / x}. That is, B contains all sets in A that do not contain themselves: For all y, ( ) y B if and only

More information

Hierarchical Design of Discrete Event Controllers: An Automated Manufacturing System Case Study

Hierarchical Design of Discrete Event Controllers: An Automated Manufacturing System Case Study Hierarchical Design of Discrete Event Controllers: An Automated Manufacturing System Case Study Technical Report Sebastian Perk, Klaus Schmidt and Thomas Moor Lehrstuhl für Regelungstechnik Friedrich-Alexander

More information

On Detectability Of Networked Discrete Event Systems

On Detectability Of Networked Discrete Event Systems Wayne State University Wayne State University Dissertations 1-1-2017 On Detectability Of Networked Discrete Event Systems Yazeed Sasi Wayne State University, Follow this and additional works at: http://digitalcommons.wayne.edu/oa_dissertations

More information

Prime Languages, Orna Kupferman, Jonathan Mosheiff. School of Engineering and Computer Science The Hebrew University, Jerusalem, Israel

Prime Languages, Orna Kupferman, Jonathan Mosheiff. School of Engineering and Computer Science The Hebrew University, Jerusalem, Israel Prime Languages, Orna Kupferman, Jonathan Mosheiff School of Engineering and Computer Science The Hebrew University, Jerusalem, Israel Abstract We say that a deterministic finite automaton (DFA) A is composite

More information

The Pumping Lemma. for all n 0, u 1 v n u 2 L (i.e. u 1 u 2 L, u 1 vu 2 L [but we knew that anyway], u 1 vvu 2 L, u 1 vvvu 2 L, etc.

The Pumping Lemma. for all n 0, u 1 v n u 2 L (i.e. u 1 u 2 L, u 1 vu 2 L [but we knew that anyway], u 1 vvu 2 L, u 1 vvvu 2 L, etc. The Pumping Lemma For every regular language L, there is a number l 1 satisfying the pumping lemma property: All w L with w l can be expressed as a concatenation of three strings, w = u 1 vu 2, where u

More information

DES. 4. Petri Nets. Introduction. Different Classes of Petri Net. Petri net properties. Analysis of Petri net models

DES. 4. Petri Nets. Introduction. Different Classes of Petri Net. Petri net properties. Analysis of Petri net models 4. Petri Nets Introduction Different Classes of Petri Net Petri net properties Analysis of Petri net models 1 Petri Nets C.A Petri, TU Darmstadt, 1962 A mathematical and graphical modeling method. Describe

More information

Büchi Automata and their closure properties. - Ajith S and Ankit Kumar

Büchi Automata and their closure properties. - Ajith S and Ankit Kumar Büchi Automata and their closure properties - Ajith S and Ankit Kumar Motivation Conventional programs accept input, compute, output result, then terminate Reactive program : not expected to terminate

More information

Achieving Fault-tolerance and Safety of Discrete-event Systems through Learning

Achieving Fault-tolerance and Safety of Discrete-event Systems through Learning 2016 American Control Conference (ACC) Boston Marriott Copley Place July 6-8, 2016. Boston, MA, USA Achieving Fault-tolerance and Safety of Discrete-event Systems through Learning Jin Dai, Ali Karimoddini,

More information

Control Synthesis of Discrete Manufacturing Systems using Timed Finite Automata

Control Synthesis of Discrete Manufacturing Systems using Timed Finite Automata Control Synthesis of Discrete Manufacturing Systems using Timed Finite utomata JROSLV FOGEL Institute of Informatics Slovak cademy of Sciences ratislav Dúbravská 9, SLOVK REPULIC bstract: - n application

More information

Obtaining the syntactic monoid via duality

Obtaining the syntactic monoid via duality Radboud University Nijmegen MLNL Groningen May 19th, 2011 Formal languages An alphabet is a non-empty finite set of symbols. If Σ is an alphabet, then Σ denotes the set of all words over Σ. The set Σ forms

More information

MODULAR SUPERVISORY CONTROL OF ASYNCHRONOUS AND HIERARCHICAL FINITE STATE MACHINES

MODULAR SUPERVISORY CONTROL OF ASYNCHRONOUS AND HIERARCHICAL FINITE STATE MACHINES MODULAR SUPERVISORY CONTROL OF ASYNCHRONOUS AND HIERARCHICAL FINITE STATE MACHINES B. Gaudin, H. Marchand VerTeCs Team, Irisa, Campus Universitaire de Beaulieu, 35042 Rennes, France E-mail: {bgaudin,hmarchan}@irisa.fr,

More information

A. Disjunctive Prognosers

A. Disjunctive Prognosers 2009 American Control Conference Hyatt Regency Riverfront, St. Louis, MO, USA June 10-12, 2009 FrB11.4 Multi-Decision Decentralized Prognosis of Failures in Discrete Event Systems Ahmed Khoumsi and Hicham

More information

Analysis and Optimization of Discrete Event Systems using Petri Nets

Analysis and Optimization of Discrete Event Systems using Petri Nets Volume 113 No. 11 2017, 1 10 ISSN: 1311-8080 (printed version); ISSN: 1314-3395 (on-line version) url: http://www.ijpam.eu ijpam.eu Analysis and Optimization of Discrete Event Systems using Petri Nets

More information

Robust Supervisory Control of a Spacecraft Propulsion System

Robust Supervisory Control of a Spacecraft Propulsion System 1 Robust Supervisory Control of a Spacecraft Propulsion System Farid Yari, Shahin Hashtrudi-Zad*, and Siamak Tafazoli In this paper the theory of supervisory control of discrete-event systems is used to

More information

Timed Automata VINO 2011

Timed Automata VINO 2011 Timed Automata VINO 2011 VeriDis Group - LORIA July 18, 2011 Content 1 Introduction 2 Timed Automata 3 Networks of timed automata Motivation Formalism for modeling and verification of real-time systems.

More information

Notes on generating functions in automata theory

Notes on generating functions in automata theory Notes on generating functions in automata theory Benjamin Steinberg December 5, 2009 Contents Introduction: Calculus can count 2 Formal power series 5 3 Rational power series 9 3. Rational power series

More information

2. Elements of the Theory of Computation, Lewis and Papadimitrou,

2. Elements of the Theory of Computation, Lewis and Papadimitrou, Introduction Finite Automata DFA, regular languages Nondeterminism, NFA, subset construction Regular Epressions Synta, Semantics Relationship to regular languages Properties of regular languages Pumping

More information

Timed Fault Tolerant Supervisory Control

Timed Fault Tolerant Supervisory Control Timed Fault Tolerant Supervisory Control TIMED FAULT TOLERANT SUPERVISORY CONTROL BY AMAL ALSUWAIDAN, B. IT. a thesis submitted to the department of computing and software and the school of graduate studies

More information

Unambiguous Morphic Images of Strings

Unambiguous Morphic Images of Strings Unambiguous Morphic Images of Strings Daniel Reidenbach, University of Kaiserslautern A joint work with: Dominik D. Freydenberger, University of Kaiserslautern Johannes C. Schneider, University of Kaiserslautern

More information

The Power of Extra Analog Neuron. Institute of Computer Science Academy of Sciences of the Czech Republic

The Power of Extra Analog Neuron. Institute of Computer Science Academy of Sciences of the Czech Republic The Power of Extra Analog Neuron Jiří Šíma Institute of Computer Science Academy of Sciences of the Czech Republic (Artificial) Neural Networks (NNs) 1. mathematical models of biological neural networks

More information

Monoidal Categories, Bialgebras, and Automata

Monoidal Categories, Bialgebras, and Automata Monoidal Categories, Bialgebras, and Automata James Worthington Mathematics Department Cornell University Binghamton University Geometry/Topology Seminar October 29, 2009 Background: Automata Finite automata

More information

Theory of Computation

Theory of Computation Thomas Zeugmann Hokkaido University Laboratory for Algorithmics http://www-alg.ist.hokudai.ac.jp/ thomas/toc/ Lecture 3: Finite State Automata Motivation In the previous lecture we learned how to formalize

More information

Hierarchical Interface-based Supervisory Control of a Flexible Manufacturing System

Hierarchical Interface-based Supervisory Control of a Flexible Manufacturing System 1 Hierarchical Interface-based Supervisory Control of a Flexible Manufacturing System R.J. Leduc, M. Lawford, and P. Dai December 22, 2005 Abstract Flexible manufacturing systems have long been touted

More information

Closure Properties of Regular Languages. Union, Intersection, Difference, Concatenation, Kleene Closure, Reversal, Homomorphism, Inverse Homomorphism

Closure Properties of Regular Languages. Union, Intersection, Difference, Concatenation, Kleene Closure, Reversal, Homomorphism, Inverse Homomorphism Closure Properties of Regular Languages Union, Intersection, Difference, Concatenation, Kleene Closure, Reversal, Homomorphism, Inverse Homomorphism Closure Properties Recall a closure property is a statement

More information

Strong Deterministic Fuzzy Automata

Strong Deterministic Fuzzy Automata Volume-5, Issue-6, December-2015 International Journal of Engineering and Management Research Page Number: 77-81 Strong Deterministic Fuzzy Automata A.Jeyanthi 1, B.Stalin 2 1 Faculty, Department of Mathematics,

More information

MOST OF the published research on control of discreteevent

MOST OF the published research on control of discreteevent IEEE TRANSACTIONS ON AUTOMATIC CONTROL, VOL. 43, NO. 1, JANUARY 1998 3 Discrete-Event Control of Nondeterministic Systems Michael Heymann and Feng Lin, Member, IEEE Abstract Nondeterminism in discrete-event

More information

Series which are both max-plus and min-plus rational are unambiguous

Series which are both max-plus and min-plus rational are unambiguous Series which are both max-plus and min-plus rational are unambiguous Sylvain Lombardy, Jean Mairesse To cite this version: Sylvain Lombardy, Jean Mairesse. Series which are both max-plus and min-plus rational

More information

Weighted Context-Free Grammars over Bimonoids

Weighted Context-Free Grammars over Bimonoids Weighted Context-Free Grammars over Bimonoids George Rahonis and Faidra Torpari Aristotle University of Thessaloniki, Greece WATA 2018 Leipzig, May 22, 2018 Faidra Torpari (Aristotle University of Thessaloniki)

More information

Failure Diagnosis of Discrete Event Systems With Linear-Time Temporal Logic Specifications

Failure Diagnosis of Discrete Event Systems With Linear-Time Temporal Logic Specifications Failure Diagnosis of Discrete Event Systems With Linear-Time Temporal Logic Specifications Shengbing Jiang and Ratnesh Kumar Abstract The paper studies failure diagnosis of discrete event systems with

More information

Application of product dioids for dead token detection in interval P-time event graphs

Application of product dioids for dead token detection in interval P-time event graphs Application of product dioids for dead token detection in interval P-time event graphs Jan Komenda Sébastien Lahaye Pavel Špaček Institute of Mathematics, Czech Academy of Sciences, Žižkova 22, 616 62

More information

arxiv: v2 [cs.fl] 29 Nov 2013

arxiv: v2 [cs.fl] 29 Nov 2013 A Survey of Multi-Tape Automata Carlo A. Furia May 2012 arxiv:1205.0178v2 [cs.fl] 29 Nov 2013 Abstract This paper summarizes the fundamental expressiveness, closure, and decidability properties of various

More information

Finite State Transducers

Finite State Transducers Finite State Transducers Eric Gribkoff May 29, 2013 Original Slides by Thomas Hanneforth (Universitat Potsdam) Outline 1 Definition of Finite State Transducer 2 Examples of FSTs 3 Definition of Regular

More information

Deterministic Finite Automata

Deterministic Finite Automata Deterministic Finite Automata COMP2600 Formal Methods for Software Engineering Katya Lebedeva Australian National University Semester 2, 2016 Slides by Ranald Clouston and Katya Lebedeva. COMP 2600 Deterministic

More information

Undecidability Results for Timed Automata with Silent Transitions

Undecidability Results for Timed Automata with Silent Transitions Fundamenta Informaticae XXI (2001) 1001 1025 1001 IOS Press Undecidability Results for Timed Automata with Silent Transitions Patricia Bouyer LSV, ENS Cachan, CNRS, France bouyer@lsv.ens-cachan.fr Serge

More information

languages by semifilter-congruences

languages by semifilter-congruences ideas Suffix semifilter-congruences Southwest Univ. Southwest Univ. Hongkong Univ. July 5 9, 2010, Nankai, China. Prefixsuffix Contents ideas 1 2 ideas 3 Suffix- 4 Prefix-suffix- Suffix Prefixsuffix ideas

More information

Finite State Automata

Finite State Automata Trento 2005 p. 1/4 Finite State Automata Automata: Theory and Practice Paritosh K. Pandya (TIFR, Mumbai, India) Unversity of Trento 10-24 May 2005 Trento 2005 p. 2/4 Finite Word Langauges Alphabet Σ is

More information

Classes and conversions

Classes and conversions Classes and conversions Regular expressions Syntax: r = ε a r r r + r r Semantics: The language L r of a regular expression r is inductively defined as follows: L =, L ε = {ε}, L a = a L r r = L r L r

More information

Fault Tolerant Controllability

Fault Tolerant Controllability 2015 American Control Conference Palmer House Hilton July 1-3, 2015. Chicago, IL, USA Fault Tolerant Controllability Simon Radel, Aos Mulahuwaish, and Ryan J. Leduc Abstract In this paper we investigate

More information

Axioms of Kleene Algebra

Axioms of Kleene Algebra Introduction to Kleene Algebra Lecture 2 CS786 Spring 2004 January 28, 2004 Axioms of Kleene Algebra In this lecture we give the formal definition of a Kleene algebra and derive some basic consequences.

More information

arxiv:cs/ v1 [cs.cc] 9 Feb 2007

arxiv:cs/ v1 [cs.cc] 9 Feb 2007 The DFAs of Finitely Different Languages Andrew Badr Ian Shipman February 1, 2008 arxiv:cs/0702053v1 [cs.cc] 9 Feb 2007 Abstract Two languages are finitely different if their symmetric difference is finite.

More information

Unary Automatic Graphs: An Algorithmic Perspective 1

Unary Automatic Graphs: An Algorithmic Perspective 1 Unary Automatic Graphs: An Algorithmic Perspective 1 This paper studies infinite graphs produced from a natural unfolding operation applied to finite graphs. Graphs produced via such operations are of

More information

Tree languages defined in first-order logic with one quantifier alternation

Tree languages defined in first-order logic with one quantifier alternation Tree languages defined in first-order logic with one quantifier alternation Miko laj Bojańczyk, Luc Segoufin Warsaw University, INRIA - LSV March 9, 2010 Abstract We study tree languages that can be defined

More information

Probabilistic Aspects of Computer Science: Probabilistic Automata

Probabilistic Aspects of Computer Science: Probabilistic Automata Probabilistic Aspects of Computer Science: Probabilistic Automata Serge Haddad LSV, ENS Paris-Saclay & CNRS & Inria M Jacques Herbrand Presentation 2 Properties of Stochastic Languages 3 Decidability Results

More information

Decentralized Failure Diagnosis of Discrete Event Systems

Decentralized Failure Diagnosis of Discrete Event Systems IEEE TRANSACTIONS ON SYSTEMS, MAN AND CYBERNETICS PART A: SYSTEMS AND HUMANS, VOL., NO., 2005 1 Decentralized Failure Diagnosis of Discrete Event Systems Wenbin Qiu, Student Member, IEEE, and Ratnesh Kumar,

More information

Universal Algebra for Logics

Universal Algebra for Logics Universal Algebra for Logics Joanna GRYGIEL University of Czestochowa Poland j.grygiel@ajd.czest.pl 2005 These notes form Lecture Notes of a short course which I will give at 1st School on Universal Logic

More information

Overlapping tile automata:

Overlapping tile automata: Overlapping tile automata: towards a language theory of overlapping structures David Janin LaBRI, Université de Bordeaux Computer Science in Russia, Ekaterinburg, june 2013 1. From strings to overlapping

More information