Inference-based Ambiguity Management in Decentralized Decision-Making: Decentralized Diagnosis of Discrete Event Systems
|
|
- Darcy Horn
- 5 years ago
- Views:
Transcription
1 Inerene-based Ambiguity Management in Deentralized Deision-Making: Deentralized Diagnosis o Disrete Event Systems Ratnesh Kumar and Shigemasa Takai Abstrat The task o deentralized deision-making involves interation o a set o loal deision-makers, eah o whih operates under limited sensing apabilities and is thus subjeted to ambiguity during the proess o deision-making. In a prior work [4] we made a key observation that suh ambiguities are o diering gradations and presented a ramework or inerening over various loal ontrol deisions o varying ambiguity levels to arrive at a global ontrol deision. We develop a similar ramework or perorming diagnosis in a deentralized setting. For eah event-trae exeuted by a system being monitored, eah loal diagnoser issues its own diagnosis deision (ailure or non-ailure or unsure, tagged with a ertain ambiguity level (zero being the minimum. A global diagnosis deision is taken to be a winning loal diagnosis deision, i.e., one with a minimum ambiguity level. The omputation o an ambiguity level or a loal deision requires an assessment o the sel-ambiguities as well as the ambiguities o the others, and an inerene based up on suh knowledge. In order to haraterize the lass o systems or whih any ault an be deteted within a uniormly bounded delay, we introdue the notion o N-inerene-diagnosability (or ailures, where the index N represents the maximum ambiguity level o any winning loal deision. We show that the odiagnosability introdued in [5] is the same as 0-inerenediagnosability; the onditional odiagnosability introdued in [13] is a type o 1-inerene-diagnosability; and the lass o higher-index inerene-diagnosable systems stritly subsumes the lass o lower-index ones I. INTRODUCTION In any deentralized deision-making paradigm, suh as deentralized ontrol or diagnosis, multiple deision-makers, eah with its limited sensing and/or ontrol apabilities, interat to ome up with the global deisions. Presene o limited sensing apabilities an lead to ambiguity in knowing the system state and thereby ambiguity in deision-making. Consider or example the problem o deentralized diagnosis [3], [9], [11], [1], [12], [5], [13] o disrete event systems (DESs. Suppose there exist two traes that are exeutable in the plant and are indistinguishable to a loal diagnoser, and one o the traes is a ailure trae while the other one is a non-ailure trae. Sine these two traes are indistinguishable, upon reeiving their observation, the loal diagnoser will be ambiguous about whether or not a ailure The researh was supported in part by the National Siene Foundation under the grants NSF-ECS , NSF-ECS , NSF-EPNES , and NSF-ECS , a DoD-EPSCoR grant through the Oie o Naval Researh under the grant N , and MEXT under Grant-in-Aid or Sientii Researh R. Kumar is with the Department o Eletrial and Computer Engineering, Iowa State University, Ames, Iowa , USA, rkumar@iastate.edu S. Takai is with the Department o Eletronis and Inormation Siene, Kyoto Institute o Tehnology, Sakyo-ku, Kyoto , Japan, takai@dj.kit.a.jp ourred. Similar situations an also arise in the setting o distributed diagnosis (i.e., one involving ommuniation among the loal diagnosers sine the problem o distributed diagnosis an be redued to an instane o deentralized diagnosis (i.e., one involving no ommuniation among the loal diagnosers [6]. Similarly, suh ambiguities an also be present in the setting o deentralized deision making or ontrol [4]. In the ontext o deentralized ontrol, a knowledge-based mehanism or assessing the sel-ambiguities was presented in [7], and later the same arhiteture was used or assessing the sel-ambiguities as well as the ambiguities o the others in [8]. The proess o utilizing the knowledge o the selambiguities together with the ambiguities o the others or the sake o deision-making was reerred to as inerening in [8] and onditioning in [14]. As is argued in [4], these prior inerening-based approahes were limited by a singlelevel o inerening, and a omprehensive ramework allowing multi-level inerening over various loal ontrol deisions o varying levels o ambiguity was irst presented in [4]. In the ontext o deentralized diagnosis, [5] suggested the ollowing simple tehnique or the management o ambiguity in the ontext o deentralized diagnosis. When a loal diagnoser is ambiguous about whether or not a ailure has ourred it simply opts to issue no diagnosis deision, i.e., a diagnosis deision is issued by a loal diagnoser only when it is unambiguous about it. This led to the introdution o the notion o odiagnosability in [5] that required that or eah ailure trae exeutable by the system being monitored, there be at least one loal diagnoser that an unambiguously determine this within a bounded number o additional transition-exeutions. An extension reported in [13] onsidered deentralized diagnosis based on the ideas o onditioning introdued in the setting o deentralized ontrol and introdued the notion o onditional odiagnosability that is weaker than odiagnosability. As is the ase with onditional oobservability, onditional odiagnosability involves a single level o inerening over the knowledge about ambiguities. In this paper we build up on the ideas o inerenebased ambiguity management in the setting o deentralized ontrol [4] and develop a ramework or inerene-based deentralized diagnosis. Our ramework supports (i inerening utilizing the knowledge o the sel-ambiguities as well as the ambiguities o the other deision makers, (ii inerening over an arbitrary number o levels o ambiguity. Eah loal diagnoser uses its observations o the system
2 behavior to ome up with its diagnosis deision together with a grade or level o ambiguity or that diagnosis deision. The omputation o an ambiguity level o a loal deision requires the assessment o the sel-ambiguities together with the ambiguities o the others. A diagnosis deision with level-zero ambiguity is taken when the loal diagnoser is unambiguous about its ailure (resp., non-ailure deision. This happens when all traes, produing the same observation as the one reeived, are ailure (resp., non-ailure traes. Otherwise, a higher ambiguity level diagnosis deision is issued. For example a ailure deision o level-one ambiguity is issued ollowing a ertain observation i there exist ertain traes, produing the same observation as the one reeived, suh that some o them are ailure traes while others are non-ailure traes. Existene o suh traes is learly a soure o ambiguity or the loal diagnoser. Yet, suppose the loal diagnoser is able to determine that or eah non-ailure trae, produing the same observation as the one reeived, there exists another loal diagnoser whih an issue a non-ailure diagnosis deision with level-zero ambiguity, then the loal diagnoser issues a ailure deision with level-one ambiguity. In general a loal diagnoser will issue a ailure (resp., nonailure deision with an ambiguity level N ollowing a ertain observation i or eah non-ailure (resp., ailure trae, produing the same observation as the one reeived, there exists another loal diagnoser that an issue a non-ailure (resp., ailure deision with an ambiguity level at most N 1. Clearly, a level-zero ambiguity diagnosis deision is based on assessment o only the sel-ambiguities, whereas a level-n ambiguity diagnosis deision is based on assessment o the sel-ambiguities together with the ambiguities o other loal diagnosers suh that or eah trae, that reates the ambiguity, there exists another loal diagnoser whih an issue a diagnosis deision with an ambiguity level at most N 1. Note in ertain situations it is possible that a loal diagnosis deision is neither ailure nor non-ailure, but unsure. As outlined above, our ramework allows inerening involving multiple-levels o ambiguities. Following the exeution o eah event all loal diagnosers reeiving a new observation issue a new diagnosis deision, tagged with a ertain level o ambiguity. The global diagnosis deision is taken to be the same as a loal diagnosis deision whose ambiguity level is the minimum. (Suh a loal deision an be onsidered to be a winning loal deision. We ormulate the notion o inerene-diagnosability to haraterize the lass o diagnosable systems in the proposed ramework o deentralized diagnosis. A system is said to be diagnosable i the global diagnosis deisions are suh that there are no missed detetions, i.e., diagnosis deision is ailure ollowing the exeution o any ailure trae within a bounded number o additional transition-exeutions, and there are no alse alarms, i.e., diagnosis deision ollowing a non-ailure (resp., ailure trae is not ailure (resp., non-ailure. For a diagnosable system, when the ambiguity level o all winning deisions are upper bounded by N, the system is said to be N-inerene-diagnosable. We study various properties o N-inerene diagnosability. In partiular we show that the notion o odiagnosability studied in [5] is the same as the notion o 0-inerene-diagnosability. Further the notion o onditional odiagnosability introdued in [13] is similar to that o 1-inerene-diagnosability. (The dierene omes beause o the use o an estimate in [13] that ignores the ourrene o unobservable events at the end. For this reason, 1-inerene-diagnosability is stronger than that the onditional odiagnosability. We also show an example that is 2-inerene-diagnosable but not onditionally diagnosable. Thus the ramework presented here subsumes and extends the ones reported earlier. We also establish that the lasses o N-inerene-diagnosable systems orm a monotonially inreasing sequene as a untion o N. We also provide an eetive test to veriy whether a given system is N- inerene-diagnosable. II. NOTATION AND PRELIMINARIES We onsider a DES modeled by a inite nondeterministi automaton G = (X, Σ, α, X 0, X m, where X is the inite set o states, Σ is the inite set o events, a partial untion α : X Σ {ε} 2 X is the transition untion, X 0 X is the set o initial states, and X m X is the set o marked or aepting states. G is said to be deterministi i the transition untion an be written as a partial untion α : X Σ X and X 0 = 1. Let Σ be the set o all inite sequenes o events inluding the empty sequene ε. Elements o Σ are alled traes, and subsets o Σ are alled languages. The transition untion α an be generalized to α : 2 X Σ 2 X in a natural way. The generated and marked (or aepted languages o G are respetively deined as, L(G := {s Σ α(x 0, s }, and L m (G := {s Σ α(x 0, s X m }. For a language K, the set o all preixes o traes in K is denoted by pr(k, i.e., pr(k = {s Σ t Σ ; st K}. K is said to be (preix-losed i K = pr(k. A language K is said to be deadlok-ree i or any s K, exists a trae t ε suh that st K; otherwise s is alled a deadloking trae o K. For eah trae s Σ, s denotes its length. For any m N, where N denotes the set o all nonnegative integers, let Σ m := {s Σ s m} and Σ m := {s Σ s m} denote the set o all traes with m or more events and m or less events respetively. III. INFERENCE-BASED DECENTRALIZED DIAGNOSIS FRAMEWORK In this paper, we onsider deentralized diagnosis where n loal diagnosers perorm the task o diagnosis without sharing their observations. We assume that the limited sensing apabilities o the ith loal diagnoser D i (i I := {1, 2,, n} an be represented as the loal observation mask, M i : Σ {ε} i {ε}, where i is the set o loally observed symbols, and M i (ε = ε. The notation Σ i,uo denotes the set o loally unobservable events, i.e., Σ i,uo = {σ Σ M i (σ = ε}.
3 Let L be a losed language representing a generated language o a plant (system to be diagnosed, and K L be a losed language representing a non-ailure speiiation language. Traes in L K are onsidered ailure traes and the task o diagnosis is to determine the exeution o any trae in L K within an additional bounded number o system exeutions. Without loss o generality, the plant language L an be taken to be deadlok-ree. Otherwise we an extend eah deadloking trae by an unbounded sequene o a newly added event that is unobservable to all diagnosers. This will make the language deadlok-ree without altering any diagnosability property sine the newly added event does not produe any observation to any o the diagnosers. Let the set C = {0, 1, φ} be the set o diagnosis deisions, where 0 represents a non-ailure deision, 1 represents a ailure deision, and φ represents an unsure deision. Eah inerene-based loal diagnoser D i is deined as a map D i : M i (L C N, where or eah s L, D i (M i (s = ( i (M i (s, n i (M i (s. Here i (M i (s C denotes the diagnosis deision o D i ollowing an observation M i (s M i (L, and n i (M i (s N denotes the ambiguity level o the diagnosis deision o D i. Let n(s be the minimal ambiguity level o loal deisions, i.e., n(s := min n i(m i (s. The deentralized diagnoser {D i } that onsists o loal diagnosers D i (i I issues global diagnosis deisions. Formally, {D i } is deined as a map {D i } : L C. For eah s L, the diagnosis deision {D i } (s is given as ollows: 0, i i I s.t. n i (M i (s = n(s; i (M i (s = 0 {D i } (s = 1, i i I s.t. n i (M i (s = n(s; i (M i (s = 1 φ, otherwise. In other words, the global diagnosis deision is taken to be the same as the minimum ambiguity level loal diagnosis deision. IV. EXISTENCE/SYNTHESIS OF INFERENCE-BASED DECENTRALIZED DIAGNOSERS Given a plant language L and a non-ailure speiiation language K L, we indutively deine a monotonially dereasing sequene o language pairs {(F k, H k } as ollows: Base step: Indution step: F 0 := L K, H 0 := K. F k+1 := F k H k+1 := H k i M i (H k i M i (F k,. The omputation o the sequene o languages (F k, H k starts with F 0 = L K, the set o Failure traes, and H 0 = K, the set o non-ailure or Healthy traes. Note that F k+1 is a sublanguage o F k onsisting o those traes or whih or eah i I there exists an M i -indistinguishable trae in H k. As a result when the plant exeutes a trae in F k+1 all the loal diagnosers will be ambiguous as to whether the exeuted trae is in F k+1 or in H k. The sublanguage H k+1 o H k an be understood in a similar ashion. Using the sequenes o language pairs (F k, H k, a loal diagnoser omputes its diagnosis deision and assoiates a level o ambiguity with suh a deision as ollows. Let N N be a given nonnegative integer. (N represents an index o diagnosability to be elaborated later. For eah s L, the ith loal diagnoser D i omputes n i (M i(s := min{n + 1, {k N M i (s / M i (H k }}, (1 n h i (M i (s := min{n + 1, {k N M i (s / M i (F k }}. (2 Note that n i (M i(s and n h i (M i(s are bounded above by N + 1. Here n i (M i(s represents the ambiguity level o a ailure deision ontemplated by the ith diagnoser ollowing the observation M i (s. When n i (M i(s < N +1, it denotes the minimum index k suh that the observation M i (s does not math with the observations o any o the traes in H k. Similarly, the notation n h i (M i(s represents the ambiguity level o a non-ailure deision ontemplated by the ith diagnoser ollowing the observation M i (s. Whih o the two ontemplated deisions is ultimately issued is deided by omparing the two ambiguity levels, n i (M i(s vs. n h i (M i(s, and avoring the smaller one. This is ormalized next. For a loal diagnoser D i : M i (L C N, its diagnosis deision and ambiguity level ollowing an observation M i (s M i (L, i.e., D i (M i (s = ( i (M i (s, n i (M i (s, is determined as ollows: and i (M i (s = 1, i n i (M i(s < n h i (M i(s 0, i n h i (M i(s < n i (M i(s φ, i n i (M i(s = n h i (M i(s (3 n i (M i (s = min{n i (M i(s, n h i (M i (s}. (4 The ollowing theorem shows that the ambiguity levels o ailure or non-ailure deisions o the deentralized diagnoser given by (1 (4 are upper bounded by N. Theorem 1: Consider the deentralized diagnoser {D i } : L C onsisting o loal diagnosers D i : M i (L C N (i I, and deined by (1 (4. Then, ( s L {D i } φ n(s N. Proo: First, we onsider the ase that {D i } = 1. For any i I suh that n(s = n i (M i (s, i (M i (s = 1. So, n(s = n i (M i(s < n h i (M i (s N + 1,
4 whih implies that n(s N. Next, we onsider the ase that {D i } = 0. For any i I suh that n(s = n i (M i (s, i (M i (s = 0. So, n(s = n h i (M i (s < n i (M i(s N + 1, whih implies that n(s N. We have the ollowing theorem whih states that there are no alse alarms under the deentralized diagnosis perormed using the loal and global diagnosers given by (1 (4. Theorem 2: Consider the deentralized diagnoser {D i } : L C onsisting o loal diagnosers D i : M i (L C N (i I, and deined by (1 (4. Then, ( s L K {D i } (s 0, (5 ( s K {D i } (s 1. (6 Proo: First, we prove (5. Suppose or ontradition that there exists s L K suh that {D i } (s = 0. Then, or any i I suh that n(s = n i (M i (s, i (M i (s = 0. Sine n(s = n i (M i (s = n h i (M i (s < n i (M i(s N + 1, n(s = n h i (M i (s = min{k N M i (s / M i (F k }, whih implies that M i (s / M i (F n(s. It ollows that s / F n(s. Sine s (L K = F 0, there exists l N suh that 0 l < n(s, s F l, and s / F l+1. So, there exists j I suh that s / j M j (H l. It ollows that M j (s / M j (H l. Sine min{k N M j (s / M j (H k } l < n(s < N + 1, n j (M j (s n j (M j(s = min{k N M j (s / M j (H k } < n(s, whih ontradits the deinition o n(s. Next, we prove (6. Suppose or ontradition that there exists s K suh that {D i } (s = 1. Then, or any i I suh that n(s = n i (M i (s, i (M i (s = 1. Sine n(s = n i (M i (s = n i (M i(s < n h i (M i (s N + 1, n(s = n i (M i(s = min{k N M i (s / M i (H k }, whih implies that M i (s / M i (H n(s. It ollows that s / H n(s. Sine s K = H 0, there exists l N suh that 0 l < n(s, s H l, and s / H l+1. So, there exists j I suh that s / j M j (F l. It ollows that M j (s / M j (F l. Sine min{k N M j (s / M j (F k } l < n(s < N + 1, n j (M j (s n h j (M j (s = min{k N M j (s / M j (F k } < n(s, whih ontradits the deinition o n(s. The task o diagnosis urther requires that there are no missed detetions or arbitrarily long ailure traes. I.e., i we let F m := (L K (L KΣ m denote ailure traes in whih a ailure ourred at least m- steps in the past, then it is desired that exists m suh that or all traes in F m, the diagnosis deision is 1, i.e., ( s F m {D i } (s = 1. (7 In the ollowing we establish a neessary and suiient ondition or (7 to hold in terms o the property o N- inerene-diagnosability or ailures. The property o N-inerene-diagnosability or ailures requires that the ailure traes that remain ambiguous ater N levels o inerring must not have inurred a ailure in a ar past (more than a bounded number o steps in the past. Deinition 1: The pair (L, K o languages is said to be N-inerene-diagnosable or ailures i there exists m N suh that F N+1 F m =. We have the ollowing lemma whih states that N- inerene-diagnosability or ailures is a suiient ondition or the nonexistene o missed ailure detetions under the deentralized diagnosis perormed using the loal and global diagnosers given by (1 (4. Lemma 1: Consider the deentralized diagnoser {D i } : L C onsisting o loal diagnosers D i : M i (L C N (i I, and deined by (1 (4. I (L, K is N-inerene-diagnosable or ailures, then {D i } satisies (7 or some m N. Proo: We onsider m N suh that F N+1 F m =. We show by ontradition that {D i } satisies (7 or m. Suppose that there exists s F m suh that {D i } (s 1. Then, there exists i I suh that n(s = n i (M i (s and i (M i (s 1. We have n(s = n h i (M i (s n i (M i(s. (8 Sine s F m and F N+1 F m =, s / F N+1. Also, s F m L K = F 0. There exists l N suh that 0 l N, s F l, and s / F l+1. So, there exists j I suh that s / j M j (H l. It ollows that M j (s / M j (H l. Sine min{k N M j (s / M j (H k } l N < N + 1, n j (M j(s = min{k N M j (s / M j (H k } < N + 1, whih implies that n(s n j (M j(s < N + 1. Sine n(s < N + 1, by (8 that n(s = n h i (M i (s = min{k N M i (s / M i (F k },
5 whih implies that M i (s / M i (F n(s. It ollows that s / F n(s. Sine s F 0, there exists l N suh that 0 l < n(s, s F l, and s / F l +1. So, there exists j I suh that s / j M j (H l. It ollows that M j (s / M j (H l. Sine min{k N M j (s / M j (H k } l < n(s < N + 1, n j (M j (s n j (M j (s = min{k N M j (s / M j (H k } < n(s, whih ontradits the deinition o n(s. The next lemma states that N-inerene-diagnosability or ailures is a neessary ondition or the nonexistene o missed ailure detetions under the deentralized diagnosis perormed using the loal and global diagnosers given by (1 (4. Lemma 2: Consider the deentralized diagnoser {D i } : L C onsisting o loal diagnosers D i : M i (L C N (i I, and deined by (1 (4. I {D i } satisies (7 or some m N, then (L, K is N-inerene-diagnosable or ailures. Proo: We onsider m N suh that {D i } satisies (7. Suppose or ontradition that F N+1 F m. We onsider any s F N+1 F m. By (7, {D i } (s = 1. It ollows rom Theorem 1 that n(s N. Then, s F N+1 F n(s+1. Sine {D i } (s = 1, there exists j I suh that n(s = n j (M j (s and j (M j (s = 1. It ollows that n(s = n j (M j(s N < N + 1. We have n(s = n j (M j(s = min{k N M j (s / M j (H k }, whih implies that M j (s / M j (H n(s. Sine s / j M j (H n(s, s / F n(s+1, whih ontradits the at that s F n(s+1. Armed with Lemmas 1 and 2, we are ready to present the main result o this setion, a theorem that proves the neessity and suiieny o N-inerene-diagnosability or ailures or the nonexistene o missed ailure detetions under the deentralized diagnosis perormed using the loal and global diagnosers given by (1 (4. Theorem 3: Consider the deentralized diagnoser {D i } : L C onsisting o loal diagnosers D i : M i (L C N (i I, and deined by (1 (4. {D i } satisies (7 or some m N i and only i (L, K is N-inerene-diagnosable or ailures. In the ollowing we present a system that is 2- inerene-diagnosable (or ailures but it is not 1-inerenediagnosable. Example 1: We onsider a plant modeled by the inite automaton G shown in Fig. 1(a, whih is a modiied version o the DES onsidered in [13]. Let n = 2, { σ, i σ {a, a M 1 (σ =,, d} ε, otherwise, { σ, i σ {b, b M 2 (σ =,, d} ε, otherwise. Also, let K L be a language generated by the inite automaton R shown in Fig. 1(b. b b a a b b a a d b a (a G d (b R a Fig. 1. Automata G and R o Example 1. We show that (L, K is 2-inerene-diagnosable or ailures. Initially, Sine F 0 = (a + b +d( + a(ε + b + b(ε + a, H 0 = pr( (ab + + ba + + d(a + + b +. M 1 (F 0 = (a + + d( + a + a, M 2 (F 0 = ( + b + d( + b + b, M 1 (H 0 = pr( (a + + a + + d(a + + +, M 2 (H 0 = pr( (b + + b + + d( + + b +, F 1 = F 0 a i M i (H 0 = (a + b + d( + a + b, ( H 1 = H 0 i M i (F 0 = pr( (a + b + d(a + + b +. It ollows that F 1 F m or any m N, whih implies that (L, K is not 0-inerene-diagnosable or ailures. Also, sine M 1 (F 1 = (a + + d( + a + ε, M 2 (F 1 = ( + b + d( + ε + b, M 1 (H 1 = pr( (a + ε + d(a + + +, M 2 (H 1 = pr( (ε + b + d( + + b +, b b
6 F 2 = F 1 i M i (H 1 = (a + b + d( + a + b, ( H 2 = H 1 i M i (F 1 = pr( (a + b + d(a + b. It ollows that F 2 F m or any m N, whih implies that (L, K is not 1-inerene-diagnosable or ailures. Moreover, sine M 1 (F 2 = (a + ε + d( + a + ε, M 2 (F 2 = (ε + b + d( + ε + b, M 1 (H 2 = pr( (a + da, M 2 (H 2 = pr( (b + db, F 3 = F 2 i M i (H 2 = (a + b + d( + a + b, ( H 3 = H 2 i M i (F 2 = H 2. We have F 2 F m = or any m 1, whih implies that (L, K is 2-inerene-diagnosable or ailures. Let N = 2. The loal deisions o D 1 and D 2 omputed using (1 (4 are shown in Table I. For example, D 1 (a is omputed as ollows. By (1 and (2, n 1 (a = 1 and n h 1(a = 2. Sine 1 = n 1 (a < nh 1(a = 2, 1 (a = 1 and n 1 (a = 1, whih implies that D 1 makes a ailure deision ollowing the observation a M 1 (L with the ambiguity level 1. TABLE I LOCAL DECISIONS OF D 1 AND D 2. t M 1 (L n 1 (t nh 1 (t 1(t n 1 (t t 3 3 φ 3 t a 3 3 φ 3 t a t a t = d 3 3 φ 3 t d t da 3 3 φ 3 t da t da t M 2 (L n 2 (t nh 2 (t 2(t n 2 (t t 3 3 φ 3 t b 3 3 φ 3 t b t b t = d 3 3 φ 3 t d t db 3 3 φ 3 t db t db Then, the global diagnosis deisions o the deentralized diagnoser {D i } are omputed as shown in Table II. For example, {D i } (a is omputed as ollows. Sine 1 = n 1 (M 1 (a < n 2 (M 2 (a = 3 and 1 (M 1 (a = 1, n(a = 1 and {D i } (a = 1. By Table II, we an veriy that {D i } satisies (5, (6, and (7 or m 1. TABLE II GLOBAL DECISIONS OF {D i }. s L n(s {D i } (s s 3 φ s a(ε + 3 φ s b(ε + 3 φ s a s b s ab 0 0 s ba 0 0 s d(ε + 3 φ s d s da(ε + 3 φ s db(ε + 3 φ s da s db s dab 0 1 s dba 0 1 Remark 1: In the system o Example 1, the event represents the ailure event. By the examples shown in [13], this system is not onditionally odiagnosable or the ailure. However as we showed above the system is 2-inerenediagnosable or the ailure. V. COMPUTATION/VERIFICATION OF INFERENCE-BASED DECENTRALIZED DIAGNOSIS/DIAGNOSERS The omputation o loal deisions using (1 (4 requires omputing the sequene o language pairs {(F k, H k }, and we present a reursive method or omputing it. Let G = (X, Σ, α, X 0, X be the plant model with L(G = L m (G = L, and R = (Y, Σ, β, Y 0, Y be a deterministi generator o the non-ailure speiiation language, i.e., L(R = L m (R = K = H 0 An aeptor or F 0 = L K is the automaton G R, where R = (Y {F }, Σ, β, Y 0, Y is R ompleted by (i adding a dump state F, and (ii adding a transition on eah event at a state in Y {F } to the dump state F i that event is not deined at that state in R. Then it an be veriied that traes in G R that end at a state with seond oordinate F belong to L K. Also, L(G R = L(G L( R = L(G Σ = L(G. Let R Fk and R Hk denote the inite aeptors o F k and H k, respetively. For eah i I, a inite aeptor o i M i (F k is onstruted as ollows: Repliate eah transition that exists in R Fk by a set o transitions on all M i -indistinguishable events. Note that sine an ε-transition is impliitly deined at eah state as a sel-loop, unobservable events will get added as sel-loops at eah state o R Fk. Then, the resulting, possibly nondeterministi, automaton M i (F k. It should be noted that this resulting automaton, denoted by i M i (R Fk, has the same state set as R Fk. In the same way, we an onstrut a inite automaton aepting i M i (H k, denoted by i M i (R Hk. Then, aepts i
7 the synhronous ompositions R Fk ( i M i (R Hk and R Hk ( i M i (R Fk aept F k+1 and H k+1, respetively. Let Y Fk and Y Hk be the state sets o R Fk and R Hk, respetively. The languages F k+1 and H k+1 are omputed rom F k and H k in O( Y Fk Y Hk I and O( Y Hk Y Fk I, respetively. The veriiation o N-inerene-diagnosability or ailures requires the heking the existene o m suh that F N+1 F m = F N+1 (L KΣ m =. For this we onsider the automaton R FN+1 G R, where R FN+1 = (Y FN+1, Σ, β FN+1, Y 0,FN+1, Y m,fn+1 is an aeptor o F N+1 obtained as outlined above, G = (X, Σ, α, X 0, X is the plant model with L(G = L m (G = L, R = (Y, Σ, β, Y 0, Y is a deterministi generator o the non-ailure speiiation language, i.e., L(R = L m (R = K, and R = (Y {F }, Σ, β, Y 0, Y is the ompletion o R by way o inlusion o an additional state F and ertain additional transitions as outlined above. Note that L(R FN+1 G R = L(R FN+1 L(G R = pr(f N+1 L(G = pr(f N+1. To veriy N-inerene-diagnosability or ailures, we veriy in the automaton R FN+1 G R whether (i the region o attration o states with the irst oordinated marked in R FN+1 (these are states reahed by traes in F N+1 ontains all states with the third oordinate F (these are states reahed by traes in L K, and (ii all strongly onneted omponents o R FN+1 G R that ontain a state with the irst oordinate marked in R FN+1 are singletons and with no sel-loops. This holds i and only i all traes in F N+1 have the property that they are inite extensions o boundary ailure traes, (L K := (L K KΣ, i.e., F N+1 (L K Σ p or some p 0, whih an be seen to be equivalent to the property o N-inerene diagnosability or ailures. Sine the above mentioned region o attration as well as the set o srongly onneted omponents an be omputed linearly in the size o R FN+1 G R [2], the omplexity o heking N-inerene-diagnosability is O( Y FN+1 X Y. VI. PROPERTIES OF N -INFERENCE-DIAGNOSABILITY In this setion, we study various properties o N-inerene diagnosable systems or ailures. First, we show that the lass o odiagnosable systems studied in [5] is equivalent to the lass o 0-inerene-diagnosable systems or ailures. Deinition 2: [5] The pair (L, K o languages is said to be odiagnosable i ( m N ( s L K( st L K s.t. t m ( i I( u E i (st; u L K, where E i (st := i M i (st L. Theorem 4: The pair (L, K o languages is 0-inerenediagnosable or ailures i and only i it is odiagnosable. Proo: ( We assume that (L, K is 0-inerenediagnosable or ailures. Then, there exists m N suh that F 1 F m =. Let s L K. Consider any st L K suh that t m. We have st (L K (L KΣ m = F m. Sine F 1 F m = and st F m F 0, st F 0 F 1, whih implies that there exists i I suh that st / i M i (H 0 = i M i (K. It ollows that M i (st / M i (K. For any u E i (st, M i (u = M i (st / M i (K, whih implies that u / K, i.e., u L K. Thus, (L, K is odiagnosable. ( We assume that (L, K is odiagnosable. Let m N be a nonnegative integer suh that the ondition o odiagnosability holds. Suppose or ontradition that F 1 F m. We onsider any s F 1 F m. Sine s F m, we an write s := tu L K where t L K and u m. Also, sine s F 1, tu i M i (K or all i I. There exists v i K L suh that M i (tu = M i (v i. Sine v i E i (tu and v i K or all i I, the ondition o odiagnosability does not hold. This is a ontradition. Next, we show that the notion o onditional F- odiagnosability introdued in [13] is similar to that o 1- inerene-diagnosability or ailures. Deinition 3: [13] The pair (L, K o languages is said to be onditionally F-odiagnosable i where ( m N ( s L K( st L K s.t. t m ENF j (u ( i I( u E i (st (u K [( j I( v ENF j (u; v K], := {s E j (u the last event o s is not in Σ i,uo }. Theorem 5: The pair (L, K o languages is 1-inerenediagnosable or ailures i and only i ( m N ( s L K( st L K s.t. t m ( i I( u E i (st (u K [( j I( v E j (u; v K]. Proo: ( We assume that (L, K is 1-inerenediagnosable or ailures. Then, there exists m N suh that F 2 F m =. Let s L K. Consider any st L K suh that t m. We have st (L K (L KΣ m = F m. Suppose or ontradition that ( i I( u i E i (st K ( j I( v ij E j (u i ; v ij L K. Sine st F 0 and st i i M i (H 0 or all i I, st F 1. Also, sine u i H 0 and u i M i (u i j M j (v ij j M j (F 0 or all j I, u i H 1. It ollows that st i M i (u i i M i (H 1 or all i I, whih implies together with st F 1 that st F 2. Thus, st F 2 F m, whih ontradits the assumption that (L, K is 1-inerene-diagnosable or ailures. ( Let m N be a nonnegative integer suh that the ondition o Theorem 5 holds. Suppose or ontradition that F 2 F m. We onsider any s F 2 F m. Sine s F m, we an write s := tu L K where t L K and u m. Also, sine s F 2, tu i M i (H 1 or all i I. There exists v i H 1 suh that M i (tu = M i (v i. It ollows that v i E i (tu and v i K or all i I. Moreover, sine
8 v i H 1, v i j M j (F 0 or all j I. There exists w ij F 0 suh that M j (v i = M j (w ij. It ollows that w ij E j (v i and w ij L K or all j I. This ontradits the ondition o Theorem 5. Sine ENF i (s E i (s or any s L and any i I, the ollowing is an immediate orollary o Theorem 5. Corollary 1: I the pair (L, K o languages is 1- inerene-diagnosable or ailures, then it is onditionally F- odiagnosable. Remark 2: Theorem 5 shows that i ENF j (u is replaed by E j (u in the deinition o onditional F-odiagnosability, then the notions o 1-inerene-diagnosability or ailures and onditional F-odiagnosability are idential. We also establish that the lasses o N-inerenediagnosable systems or ailures orm a monotonially inreasing sequene as a untion o N. Sine the sequene o language pairs {(F k, H k } is monotonially dereasing or any k N, the ollowing result is easily obtained (the proo is omitted. Theorem 6: For any N N, i the pair (L, K o languages is N-inerene-diagnosable or ailures, then it is (N + 1-inerene-diagnosable or ailures. The onverse relation o Theorem 6 need not hold. For example, the system o Example 1 is 2-inerene-diagnosable or ailures, but not 1-inerene-diagnosable. VII. CONCLUSION A key issue in deentralized deision-making is the usion o the loal deisions to arrive at a global deision. A key observation we made in a prior work [4] is that suh deision usion an be ailitated by assessing ambiguity levels o eah loal deision-maker (arising due to its limited sensing apability and using that knowledge to arrive at a global deision. Our prior work [4] used this idea to propose an inerene-based ramework or deentralized ontrol (inerening over sel-ambiguities and those o others is used or the assessment o the ambiguity level o ones own deision. The present paper proposes an inerene-based ramework or deentralized diagnosis. The proposed ramework is able to extend the prior approahes to deentralized diagnosis (suh as those reported in [5], [13]. Also through the work reported in [6] it is evident that a deentralized diagnosis ramework (one involving no ommuniation among loal diagnosers is also useul or distributed diagnosis appliations (one involving ommuniation among loal diagnosers, and so the appliability o the proposed ramework extends to the setting o distributed diagnosis. It should be noted that as the order o inerening inorporated into deentralized deision-making is enhaned, the orresponding ost o omputing the loal deisions is also inreased (as ormalized in Setion V. So the additional gain resulting rom a higher order o inerening omes at an additional omputational ost. REFERENCES [1] R. K. Boel and J. H. van Shuppen. Deentralized ailure diagnosis or disrete-event systems with onstrained ommuniation between diagnosers. In Proeedings o International Workshop on Disrete Event Systems, [2] Y. Brave and M. Heymann. On stabilization o disrete event proesses. International Journal o Control, 51: , [3] R. Debouk, S. Laortune, and D. Teneketzis. Coordinated deentralized protools or ailure diagnosis o disrete event systems. Disrete Event Dynamial Systems: Theory and Appliations, 10:33 79, [4] R. Kumar and S. Takai. Inerene-based ambiguity management in deentralized deision-making: Deentralized ontrol o disrete event systems. In Proeeding o 2005 IEEE Conerene on Deision and Control and European Control Conerene, Seville, Spain, Deember [5] W. Qiu and R. Kumar. Deentralized ailure diagnosis o disrete event systems. In Proeedings o 2004 International Workshop on Disrete Event Systems, Reims, Frane, September [6] W. Qiu and R. Kumar. Distributed ailure diagnosis under bounded delay using immediate observation passing protool. In Proeedings o 2005 Amerian Control Conerene, Portland, OR, June [7] S. L. Riker and K. Rudie. Know means no: Inorporating knowledge into disrete-event ontrol systems. IEEE Transations on Automati Control, 45: , September [8] S. L. Riker and K. Rudie. Knowledge is a terrible thing to waste: Using inerene in disrete-event ontrol problems. In Proeedings o 2003 Amerian Control Conerene, pages , Denver, CO, [9] S. L. Riker and J. H. van Shuppen. Deentralized ailure diagnosis with asynhronous ommuniation between supervisors. In Proeedings o the European Control Conerene, pages , [10] M. Sampath, R. Sengupta, S. Laortune, K. Sinaamohideen, and D. Teneketzis. Diagnosability o disrete-event systems. IEEE Transations on Automati Control, 40(9: , [11] R. Sengupta and S. Tripakis. Deentralized diagnosis o regular language is undeidable. In Proeedings o IEEE Conerene on Deision and Control, pages , Las Vegas, NV, Deember [12] R. Su, W. M. Wonham, J. Kurien, and X. Koutsoukos. Distributed diagnosis or qualitative systems. In Proeedings o International Workshop on Disrete Event Systems, [13] Y. Wang, T.-S. Yoo, and S. Laortune. New results on deentralized diagnosis o disrete-event systems. In Proeedings o 2004 Annual Allerton Conerene, [14] T.-S. Yoo and S. Laortune. Deentralized supervisory ontrol with onditional deisions: Supervisor existene. IEEE Transations on Automati Control, 49(11: , 2004.
Nonreversibility of Multiple Unicast Networks
Nonreversibility of Multiple Uniast Networks Randall Dougherty and Kenneth Zeger September 27, 2005 Abstrat We prove that for any finite direted ayli network, there exists a orresponding multiple uniast
More informationON THE FOUR-COLOUR CONJECTURE. By W. T. TUTTE. [Received 27 November 1945 Read 13 December 1945]
ON THE FOUR COLOUR CONJECTURE 37 ON THE FOUR-COLOUR CONJECTURE By W. T. TUTTE [Reeived 7 November 945 Read 3 Deember 945]. Introdution The maps disussed in this paper are dissetions o suraes into simple
More informationParallel disrete-event simulation is an attempt to speed-up the simulation proess through the use of multiple proessors. In some sense parallel disret
Exploiting intra-objet dependenies in parallel simulation Franeso Quaglia a;1 Roberto Baldoni a;2 a Dipartimento di Informatia e Sistemistia Universita \La Sapienza" Via Salaria 113, 198 Roma, Italy Abstrat
More informationDECENTRALIZED 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 informationDiagnosis of Dense-Time Systems using Digital-Clocks
Diagnosis of Dense-Time Systems using Digital-Clocks Shengbing Jiang GM R&D and Planning Mail Code 480-106-390 Warren, MI 48090-9055 Email: shengbing.jiang@gm.com Ratnesh Kumar Dept. of Elec. & Comp. Eng.
More informationFinite Formulation of Electromagnetic Field
Finite Formulation o Eletromagneti Field Enzo TONTI Dept.Civil Engin., Univ. o Trieste, Piazzale Europa 1, 34127 Trieste, Italia. e-mail: tonti@univ.trieste.it Otober 16, 2000 Abstrat This paper shows
More informationLearning to model sequences generated by switching distributions
earning to model sequenes generated by swithing distributions Yoav Freund A Bell abs 00 Mountain Ave Murray Hill NJ USA Dana on omputer Siene nstitute Hebrew University Jerusalem srael Abstrat We study
More information(q) -convergence. Comenius University, Bratislava, Slovakia
Annales Mathematiae et Informatiae 38 (2011) pp. 27 36 http://ami.ektf.hu On I (q) -onvergene J. Gogola a, M. Mačaj b, T. Visnyai b a University of Eonomis, Bratislava, Slovakia e-mail: gogola@euba.sk
More informationOn Component Order Edge Reliability and the Existence of Uniformly Most Reliable Unicycles
Daniel Gross, Lakshmi Iswara, L. William Kazmierzak, Kristi Luttrell, John T. Saoman, Charles Suffel On Component Order Edge Reliability and the Existene of Uniformly Most Reliable Uniyles DANIEL GROSS
More informationEigenvalues of tridiagonal matrix using Strum Sequence and Gerschgorin theorem
Eigenvalues o tridiagonal matrix using Strum Sequene and Gershgorin theorem T.D.Roopamala Department o Computer Siene and Engg., Sri Jayahamarajendra College o Engineering Mysore INDIA roopa_td@yahoo.o.in
More informationETNA Kent State University
% { Eletroni Transations on Numerial Analysis. Volume 21, pp. 20-27, 2005. Copyright 2005,. ISSN 1068-9613. QR FACTORIZATIONS USING A RESTRICTED SET OF ROTATIONS DIANNE P. O LEARY AND STEPHEN S. BULLOCK
More informationSufficient Conditions for a Flexible Manufacturing System to be Deadlocked
Paper 0, INT 0 Suffiient Conditions for a Flexile Manufaturing System to e Deadloked Paul E Deering, PhD Department of Engineering Tehnology and Management Ohio University deering@ohioedu Astrat In reent
More informationarxiv: v2 [cs.dm] 4 May 2018
Disrete Morse theory for the ollapsibility of supremum setions Balthazar Bauer INRIA, DIENS, PSL researh, CNRS, Paris, Frane Luas Isenmann LIRMM, Université de Montpellier, CNRS, Montpellier, Frane arxiv:1803.09577v2
More informationStructural Reconfiguration of Systems under Behavioral Adaptation
Strutural Reonfiguration of Systems under Behavioral Adaptation Carlos Canal a, Javier Cámara b, Gwen Salaün a Department of Computer Siene, University of Málaga, Spain b Department of Informatis Engineering,
More informationSearching All Approximate Covers and Their Distance using Finite Automata
Searhing All Approximate Covers and Their Distane using Finite Automata Ondřej Guth, Bořivoj Melihar, and Miroslav Balík České vysoké učení tehniké v Praze, Praha, CZ, {gutho1,melihar,alikm}@fel.vut.z
More informationComplexity of Regularization RBF Networks
Complexity of Regularization RBF Networks Mark A Kon Department of Mathematis and Statistis Boston University Boston, MA 02215 mkon@buedu Leszek Plaskota Institute of Applied Mathematis University of Warsaw
More informationPacking Plane Spanning Trees into a Point Set
Paking Plane Spanning Trees into a Point Set Ahmad Biniaz Alfredo Garía Abstrat Let P be a set of n points in the plane in general position. We show that at least n/3 plane spanning trees an be paked into
More informationDecentralized Diagnosis for Nonfailures of Discrete Event Systems Using Inference-Based Ambiguity Management
IEEE TRANSACTIONS ON SYSTEMS, MAN, AND CYBERNETICS PART A: SYSTEMS AND HUMANS, VOL. XX, NO. X, XXX 2009 1 Deentrlized Dignosis or Nonilures o Disrete Event Systems Using Inerene-Bsed Amiguity Mngement
More informationPREDICTION OF CONCRETE COMPRESSIVE STRENGTH
PREDICTION OF CONCRETE COMPRESSIVE STRENGTH Dunja Mikuli (1), Ivan Gabrijel (1) and Bojan Milovanovi (1) (1) Faulty o Civil Engineering, University o Zagreb, Croatia Abstrat A ompressive strength o onrete
More informationA Functional Representation of Fuzzy Preferences
Theoretial Eonomis Letters, 017, 7, 13- http://wwwsirporg/journal/tel ISSN Online: 16-086 ISSN Print: 16-078 A Funtional Representation of Fuzzy Preferenes Susheng Wang Department of Eonomis, Hong Kong
More informationAcoustic Attenuation Performance of Helicoidal Resonator Due to Distance Change from Different Cross-sectional Elements of Cylindrical Ducts
Exerpt rom the Proeedings o the COMSOL Conerene 1 Paris Aousti Attenuation Perormane o Helioidal Resonator Due to Distane Change rom Dierent Cross-setional Elements o Cylindrial Duts Wojieh ŁAPKA* Division
More informationSensitivity Analysis in Markov Networks
Sensitivity Analysis in Markov Networks Hei Chan and Adnan Darwihe Computer Siene Department University of California, Los Angeles Los Angeles, CA 90095 {hei,darwihe}@s.ula.edu Abstrat This paper explores
More informationDynamic Programming and Multi Objective Linear Programming approaches
0 NSP Vol. () (0), 9 Dynami Programming and Multi Objetive Linear Programming approahes P. K. De and Amita Bhinher Department o Mathematis, National Institute o Tehnology SILCHAR - 788 00 (Assam) India
More informationarxiv: v2 [cs.oh] 25 Apr 2013
Online Energy Generation Sheduling or Mirogrids with Intermittent Energy Soures and Co-Generation Lian Lu arxiv:1211.4473v2 [s.oh] 25 Apr 213 The Department o Inormation Engineering The Chinese University
More informationHankel Optimal Model Order Reduction 1
Massahusetts Institute of Tehnology Department of Eletrial Engineering and Computer Siene 6.245: MULTIVARIABLE CONTROL SYSTEMS by A. Megretski Hankel Optimal Model Order Redution 1 This leture overs both
More informationA Recursive Approach to the Kauffman Bracket
Applied Mathematis, 204, 5, 2746-2755 Published Online Otober 204 in SiRes http://wwwsirporg/journal/am http://ddoiorg/04236/am20457262 A Reursive Approah to the Kauffman Braet Abdul Rauf Nizami, Mobeen
More informationControl of industrial robots. Control of the interaction
Control o industrial robots Control o the interation Pro. Paolo Roo (paolo.roo@polimi.it) Politenio di Milano Dipartimento di Elettronia, Inormazione e Bioingegneria Introdution So ar we have assumed that
More informationFormal Specification for Transportation Cyber Physical Systems
Formal Speifiation for Transportation Cyber Physial Systems ihen Zhang, Jifeng He and Wensheng Yu Shanghai Key aboratory of Trustworthy Computing East China Normal University Shanghai 200062, China Zhanglihen1962@163.om
More informationExploring the feasibility of on-site earthquake early warning using close-in records of the 2007 Noto Hanto earthquake
Exploring the feasibility of on-site earthquake early warning using lose-in reords of the 2007 Noto Hanto earthquake Yih-Min Wu 1 and Hiroo Kanamori 2 1. Department of Geosienes, National Taiwan University,
More informationAssessing the Performance of a BCI: A Task-Oriented Approach
Assessing the Performane of a BCI: A Task-Oriented Approah B. Dal Seno, L. Mainardi 2, M. Matteui Department of Eletronis and Information, IIT-Unit, Politenio di Milano, Italy 2 Department of Bioengineering,
More informationA Spatiotemporal Approach to Passive Sound Source Localization
A Spatiotemporal Approah Passive Sound Soure Loalization Pasi Pertilä, Mikko Parviainen, Teemu Korhonen and Ari Visa Institute of Signal Proessing Tampere University of Tehnology, P.O.Box 553, FIN-330,
More informationDiagnosing Hybrid Systems: a Bayesian Model Selection Approach
Diagnosing Hybrid Systems: a Bayesian Model Seletion Approah Sheila A MIlraith Knowledge Systems Laboratory Stanord University Stanord CA 94025 Phone: 650-723-7932, Fax: 650-725-5850 sam@kslstanordedu
More informationBayesian Optimization Under Uncertainty
Bayesian Optimization Under Unertainty Justin J. Beland University o Toronto justin.beland@mail.utoronto.a Prasanth B. Nair University o Toronto pbn@utias.utoronto.a Abstrat We onsider the problem o robust
More informationProbabilistic and nondeterministic aspects of Anonymity 1
MFPS XX1 Preliminary Version Probabilisti and nondeterministi aspets of Anonymity 1 Catusia Palamidessi 2 INRIA and LIX Éole Polytehnique, Rue de Salay, 91128 Palaiseau Cedex, FRANCE Abstrat Anonymity
More informationLightpath routing for maximum reliability in optical mesh networks
Vol. 7, No. 5 / May 2008 / JOURNAL OF OPTICAL NETWORKING 449 Lightpath routing for maximum reliability in optial mesh networks Shengli Yuan, 1, * Saket Varma, 2 and Jason P. Jue 2 1 Department of Computer
More informationComputer Science 786S - Statistical Methods in Natural Language Processing and Data Analysis Page 1
Computer Siene 786S - Statistial Methods in Natural Language Proessing and Data Analysis Page 1 Hypothesis Testing A statistial hypothesis is a statement about the nature of the distribution of a random
More informationDIGITAL DISTANCE RELAYING SCHEME FOR PARALLEL TRANSMISSION LINES DURING INTER-CIRCUIT FAULTS
CHAPTER 4 DIGITAL DISTANCE RELAYING SCHEME FOR PARALLEL TRANSMISSION LINES DURING INTER-CIRCUIT FAULTS 4.1 INTRODUCTION Around the world, environmental and ost onsiousness are foring utilities to install
More informationOptimal Control of Air Pollution
Punjab University Journal of Mathematis (ISSN 1016-2526) Vol. 49(1)(2017) pp. 139-148 Optimal Control of Air Pollution Y. O. Aderinto and O. M. Bamigbola Mathematis Department, University of Ilorin, Ilorin,
More informationReliability Estimation of Solder Joints Under Thermal Fatigue with Varying Parameters by using FORM and MCS
Proeedings o the World Congress on Engineering 2007 Vol II Reliability Estimation o Solder Joints Under Thermal Fatigue with Varying Parameters by using FORM and MCS Ouk Sub Lee, Yeon Chang Park, and Dong
More informationStochastic Adaptive Learning Rate in an Identification Method: An Approach for On-line Drilling Processes Monitoring
9 Amerian Control Conerene Hyatt Regeny Riverront, St. Louis, MO, USA June -, 9 FrB3. Stohasti Adaptive Learning Rate in an Identiiation Method: An Approah or On-line Drilling Proesses Monitoring A. BA,
More informationf 2 f n where m is the total mass of the object. Expression (6a) is plotted in Figure 8 for several values of damping ( ).
F o F o / k A = = 6 k 1 + 1 + n r n n n RESONANCE It is seen in Figure 7 that displaement and stress levels tend to build up greatly when the oring requeny oinides with the natural requeny, the buildup
More informationTight bounds for selfish and greedy load balancing
Tight bounds for selfish and greedy load balaning Ioannis Caragiannis Mihele Flammini Christos Kaklamanis Panagiotis Kanellopoulos Lua Mosardelli Deember, 009 Abstrat We study the load balaning problem
More informationDecentralized 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 informationRemark 4.1 Unlike Lyapunov theorems, LaSalle s theorem does not require the function V ( x ) to be positive definite.
Leture Remark 4.1 Unlike Lyapunov theorems, LaSalle s theorem does not require the funtion V ( x ) to be positive definite. ost often, our interest will be to show that x( t) as t. For that we will need
More informationA novel Infrared Thermography (IRT) based experimental technique for distributed temperature measurements in hot gas flows
QIRT 1 1 http://dx.doi.org/1.1611/qirt.1.45 th International Conerene on Quantitative InraRed Thermography July 7-3, 1, Québe (Canada A novel Inrared Thermography (IRT based experimental tehnique or distributed
More informationMethods of evaluating tests
Methods of evaluating tests Let X,, 1 Xn be i.i.d. Bernoulli( p ). Then 5 j= 1 j ( 5, ) T = X Binomial p. We test 1 H : p vs. 1 1 H : p>. We saw that a LRT is 1 if t k* φ ( x ) =. otherwise (t is the observed
More informationPushdown Specifications
Orna Kupferman Hebrew University Pushdown Speifiations Nir Piterman Weizmann Institute of Siene une 9, 2002 Moshe Y Vardi Rie University Abstrat Traditionally, model heking is applied to finite-state systems
More informationOrdered fields and the ultrafilter theorem
F U N D A M E N T A MATHEMATICAE 59 (999) Ordered fields and the ultrafilter theorem by R. B e r r (Dortmund), F. D e l o n (Paris) and J. S h m i d (Dortmund) Abstrat. We prove that on the basis of ZF
More informationSome GIS Topological Concepts via Neutrosophic Crisp Set Theory
New Trends in Neutrosophi Theory and Appliations A.A.SALAMA, I.M.HANAFY, HEWAYDA ELGHAWALBY 3, M.S.DABASH 4,2,4 Department of Mathematis and Computer Siene, Faulty of Sienes, Port Said University, Egypt.
More informationDiscrete Bessel functions and partial difference equations
Disrete Bessel funtions and partial differene equations Antonín Slavík Charles University, Faulty of Mathematis and Physis, Sokolovská 83, 186 75 Praha 8, Czeh Republi E-mail: slavik@karlin.mff.uni.z Abstrat
More information15.12 Applications of Suffix Trees
248 Algorithms in Bioinformatis II, SoSe 07, ZBIT, D. Huson, May 14, 2007 15.12 Appliations of Suffix Trees 1. Searhing for exat patterns 2. Minimal unique substrings 3. Maximum unique mathes 4. Maximum
More informationProbabilistic Simulation Approach to Evaluate the Tooth-Root Strength of Spur Gears with FEM-Based Verification
Engineering, 2011, 3, 1137-1148 doi:10.4236/eng.2011.312142 Published Online Deember 2011 (http://www.sirp.org/journal/eng) Abstrat Probabilisti Simulation Approah to Evaluate the Tooth-Root Strength o
More informationSupervision Patterns in Discrete Event Systems Diagnosis
Supervision Patterns in Discrete Event Systems Diagnosis Thierry Jéron, Hervé Marchand, Sophie Pinchinat, Marie-Odile Cordier IRISA, Campus Universitaire de Beaulieu, 35042 Rennes, rance {irstame.ame}@irisa.r
More informationCase I: 2 users In case of 2 users, the probability of error for user 1 was earlier derived to be 2 A1
MUTLIUSER DETECTION (Letures 9 and 0) 6:33:546 Wireless Communiations Tehnologies Instrutor: Dr. Narayan Mandayam Summary By Shweta Shrivastava (shwetash@winlab.rutgers.edu) bstrat This artile ontinues
More informationarxiv:nucl-th/ v1 27 Jul 1999
Eetive Widths and Eetive Number o Phonons o Multiphonon Giant Resonanes L.F. Canto, B.V. Carlson, M.S. Hussein 3 and A.F.R. de Toledo Piza 3 Instituto de Físia, Universidade do Rio de Janeiro, CP 6858,
More informationChapter 8 Hypothesis Testing
Leture 5 for BST 63: Statistial Theory II Kui Zhang, Spring Chapter 8 Hypothesis Testing Setion 8 Introdution Definition 8 A hypothesis is a statement about a population parameter Definition 8 The two
More informationMillennium Relativity Acceleration Composition. The Relativistic Relationship between Acceleration and Uniform Motion
Millennium Relativity Aeleration Composition he Relativisti Relationship between Aeleration and niform Motion Copyright 003 Joseph A. Rybzyk Abstrat he relativisti priniples developed throughout the six
More informationA Characterization of Wavelet Convergence in Sobolev Spaces
A Charaterization of Wavelet Convergene in Sobolev Spaes Mark A. Kon 1 oston University Louise Arakelian Raphael Howard University Dediated to Prof. Robert Carroll on the oasion of his 70th birthday. Abstrat
More informationError Bounds for Context Reduction and Feature Omission
Error Bounds for Context Redution and Feature Omission Eugen Bek, Ralf Shlüter, Hermann Ney,2 Human Language Tehnology and Pattern Reognition, Computer Siene Department RWTH Aahen University, Ahornstr.
More informationDanielle Maddix AA238 Final Project December 9, 2016
Struture and Parameter Learning in Bayesian Networks with Appliations to Prediting Breast Caner Tumor Malignany in a Lower Dimension Feature Spae Danielle Maddix AA238 Final Projet Deember 9, 2016 Abstrat
More informationFrugality Ratios And Improved Truthful Mechanisms for Vertex Cover
Frugality Ratios And Improved Truthful Mehanisms for Vertex Cover Edith Elkind Hebrew University of Jerusalem, Israel, and University of Southampton, Southampton, SO17 1BJ, U.K. Leslie Ann Goldberg University
More informationDecentralized 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 informationDoppler effect of the rupture process of the great M W 7.9 Wenchuan earthquake
Earthq Si (1)3: 535 539 535 539 Doi: 1.17/s11589-1-75-4 Doppler eet o the rupture proess o the great M W 7.9 Wenhuan earthquake Ge Jin 1,, Youai Tang 1 Shiyong Zhou 1 and Yongshun John Chen 1 1 Institute
More informationFair Integrated Scheduling of Soft Real-time Tardiness Classes on Multiprocessors
Fair Integrated Sheduling of Soft Real-time Tardiness Classes on Multiproessors UmaMaheswari C. Devi and James H. Anderson Department of Computer Siene, The University of North Carolina, Chapel Hill, NC
More informationControl Theory association of mathematics and engineering
Control Theory assoiation of mathematis and engineering Wojieh Mitkowski Krzysztof Oprzedkiewiz Department of Automatis AGH Univ. of Siene & Tehnology, Craow, Poland, Abstrat In this paper a methodology
More informationAverage Rate Speed Scaling
Average Rate Speed Saling Nikhil Bansal David P. Bunde Ho-Leung Chan Kirk Pruhs May 2, 2008 Abstrat Speed saling is a power management tehnique that involves dynamially hanging the speed of a proessor.
More informationPhysical Laws, Absolutes, Relative Absolutes and Relativistic Time Phenomena
Page 1 of 10 Physial Laws, Absolutes, Relative Absolutes and Relativisti Time Phenomena Antonio Ruggeri modexp@iafria.om Sine in the field of knowledge we deal with absolutes, there are absolute laws that
More informationArray Design for Superresolution Direction-Finding Algorithms
Array Design for Superresolution Diretion-Finding Algorithms Naushad Hussein Dowlut BEng, ACGI, AMIEE Athanassios Manikas PhD, DIC, AMIEE, MIEEE Department of Eletrial Eletroni Engineering Imperial College
More informationDevelopment of moving sound source localization system
Eletroni Journal «Tehnial Aoustis» http://www.ejta.org 2006, 11 Doh-Hyoung Kim 1, Youngjin Park 2 Korea Advaned Institute o Siene and Tehnology, ME3076, 373-1, Guseong-Dong, Yuseong-Gu, Daejon, 305-701,
More informationarxiv: v2 [math.pr] 9 Dec 2016
Omnithermal Perfet Simulation for Multi-server Queues Stephen B. Connor 3th Deember 206 arxiv:60.0602v2 [math.pr] 9 De 206 Abstrat A number of perfet simulation algorithms for multi-server First Come First
More informationInvestigation on entry capacities of single-lane roundabouts
University o Wollongong Researh Online Faulty o Engineering and Inormation Sienes - Papers: Part A Faulty o Engineering and Inormation Sienes 014 Investigation on entry apaities o single-lane roundabouts
More informationNormative and descriptive approaches to multiattribute decision making
De. 009, Volume 8, No. (Serial No.78) China-USA Business Review, ISSN 57-54, USA Normative and desriptive approahes to multiattribute deision making Milan Terek (Department of Statistis, University of
More informationOptimization of Statistical Decisions for Age Replacement Problems via a New Pivotal Quantity Averaging Approach
Amerian Journal of heoretial and Applied tatistis 6; 5(-): -8 Published online January 7, 6 (http://www.sienepublishinggroup.om/j/ajtas) doi:.648/j.ajtas.s.65.4 IN: 36-8999 (Print); IN: 36-96 (Online)
More informationONLINE APPENDICES for Cost-Effective Quality Assurance in Crowd Labeling
ONLINE APPENDICES for Cost-Effetive Quality Assurane in Crowd Labeling Jing Wang Shool of Business and Management Hong Kong University of Siene and Tehnology Clear Water Bay Kowloon Hong Kong jwang@usthk
More informationON THE VARIATION OF THE HARDY LITTLEWOOD MAXIMAL FUNCTION
Annales Aademiæ Sientiarum Fenniæ Mathematia Volumen 40, 2015, 109 133 ON THE VARIATION OF THE HARDY LITTLEWOOD MAXIMAL FUNCTION Ondřej Kurka Charles University, Faulty o Mathematis Physis, Department
More informationGeneral Closed-form Analytical Expressions of Air-gap Inductances for Surfacemounted Permanent Magnet and Induction Machines
General Closed-form Analytial Expressions of Air-gap Indutanes for Surfaemounted Permanent Magnet and Indution Mahines Ronghai Qu, Member, IEEE Eletroni & Photoni Systems Tehnologies General Eletri Company
More informationOptimal Distributed Estimation Fusion with Transformed Data
Optimal Distribute Estimation Fusion with Transforme Data Zhansheng Duan X. Rong Li Department of Eletrial Engineering University of New Orleans New Orleans LA 70148 U.S.A. Email: {zuanxli@uno.eu Abstrat
More informationProduct Policy in Markets with Word-of-Mouth Communication. Technical Appendix
rodut oliy in Markets with Word-of-Mouth Communiation Tehnial Appendix August 05 Miro-Model for Inreasing Awareness In the paper, we make the assumption that awareness is inreasing in ustomer type. I.e.,
More informationJournal of Inequalities in Pure and Applied Mathematics
Journal of Inequalities in Pure and Applied Mathematis A NEW ARRANGEMENT INEQUALITY MOHAMMAD JAVAHERI University of Oregon Department of Mathematis Fenton Hall, Eugene, OR 97403. EMail: javaheri@uoregon.edu
More informationmax min z i i=1 x j k s.t. j=1 x j j:i T j
AM 221: Advaned Optimization Spring 2016 Prof. Yaron Singer Leture 22 April 18th 1 Overview In this leture, we will study the pipage rounding tehnique whih is a deterministi rounding proedure that an be
More informationSensitivity of Spectrum Sensing Techniques to RF impairments
Sensitivity of Spetrum Sensing Tehniques to RF impairments Jonathan Verlant-Chenet Julien Renard Jean-Mihel Driot Philippe De Donker François Horlin Université Libre de Bruelles - OPERA Dpt., Avenue F.D.
More informationThe Impact of Time on the Session Problem Injong Rhee Jennifer L. Welch Department of Computer Science CB 3175 Sitterson Hall University of North Caro
The Impat of Time on the Session Problem Injong Rhee Jennifer L. Welh Department of Computer Siene CB 3175 Sitterson Hall University of North Carolina at Chapel Hill Chapel Hill, N.C. 27599-3175 Abstrat
More informationStability of Hybrid Automata with Average Dwell Time: An Invariant Approach
1 Stability of Hybrid Automata with Average Dwell Time An Invariant Approah Sayan Mitra* Computer Siene and Artifiial Intelligene Laboratory Massahusetts Institute of Tehnology 200 Tehnology Square Cambridge,
More informationSensor management for PRF selection in the track-before-detect context
Sensor management for PRF seletion in the tra-before-detet ontext Fotios Katsilieris, Yvo Boers, and Hans Driessen Thales Nederland B.V. Haasbergerstraat 49, 7554 PA Hengelo, the Netherlands Email: {Fotios.Katsilieris,
More informationOn the Licensing of Innovations under Strategic Delegation
On the Liensing of Innovations under Strategi Delegation Judy Hsu Institute of Finanial Management Nanhua University Taiwan and X. Henry Wang Department of Eonomis University of Missouri USA Abstrat This
More informationAn Integrated Architecture of Adaptive Neural Network Control for Dynamic Systems
An Integrated Arhiteture of Adaptive Neural Network Control for Dynami Systems Robert L. Tokar 2 Brian D.MVey2 'Center for Nonlinear Studies, 2Applied Theoretial Physis Division Los Alamos National Laboratory,
More informationA Queueing Model for Call Blending in Call Centers
A Queueing Model for Call Blending in Call Centers Sandjai Bhulai and Ger Koole Vrije Universiteit Amsterdam Faulty of Sienes De Boelelaan 1081a 1081 HV Amsterdam The Netherlands E-mail: {sbhulai, koole}@s.vu.nl
More informationI F I G R e s e a r c h R e p o r t. Minimal and Hyper-Minimal Biautomata. IFIG Research Report 1401 March Institut für Informatik
I F I G R e s e a r h R e p o r t Institut für Informatik Minimal and Hyper-Minimal Biautomata Markus Holzer Seastian Jakoi IFIG Researh Report 1401 Marh 2014 Institut für Informatik JLU Gießen Arndtstraße
More informationFiber Optic Cable Transmission Losses with Perturbation Effects
Fiber Opti Cable Transmission Losses with Perturbation Effets Kampanat Namngam 1*, Preeha Yupapin 2 and Pakkinee Chitsakul 1 1 Department of Mathematis and Computer Siene, Faulty of Siene, King Mongkut
More informationDiscrete Generalized Burr-Type XII Distribution
Journal of Modern Applied Statistial Methods Volume 13 Issue 2 Artile 13 11-2014 Disrete Generalized Burr-Type XII Distribution B. A. Para University of Kashmir, Srinagar, India, parabilal@gmail.om T.
More informationSynthesis of verifiably hazard-free asynchronous control circuits
Synthesis of verifiably hazardfree asynhronous ontrol iruits L. Lavagno Dept. of EECS University of California, Berkeley K. Keutzer AT&T Bell Laboratories Murray Hill, NJ November 9, 990 A. SangiovanniVinentelli
More informationFrequency hopping does not increase anti-jamming resilience of wireless channels
Frequeny hopping does not inrease anti-jamming resiliene of wireless hannels Moritz Wiese and Panos Papadimitratos Networed Systems Seurity Group KTH Royal Institute of Tehnology, Stoholm, Sweden {moritzw,
More informationGeneral Equilibrium. What happens to cause a reaction to come to equilibrium?
General Equilibrium Chemial Equilibrium Most hemial reations that are enountered are reversible. In other words, they go fairly easily in either the forward or reverse diretions. The thing to remember
More informationChapter 3 Church-Turing Thesis. CS 341: Foundations of CS II
CS 341: Foundations of CS II Marvin K. Nakayama Computer Siene Department New Jersey Institute of Tehnology Newark, NJ 07102 CS 341: Chapter 3 3-2 Contents Turing Mahines Turing-reognizable Turing-deidable
More informationRobust Flight Control Design for a Turn Coordination System with Parameter Uncertainties
Amerian Journal of Applied Sienes 4 (7): 496-501, 007 ISSN 1546-939 007 Siene Publiations Robust Flight ontrol Design for a urn oordination System with Parameter Unertainties 1 Ari Legowo and Hiroshi Okubo
More informationMotor Sizing Application Note
PAE-TILOGY Linear Motors 70 Mill orest d. Webster, TX 77598 (8) 6-7750 ax (8) 6-7760 www.trilogysystems.om E-mail emn_support_trilogy@parker.om Motor Sizing Appliation Note By Jak Marsh Introdution Linear
More informationSome Properties on Nano Topology Induced by Graphs
AASCIT Journal of anosiene 2017; 3(4): 19-23 http://wwwaasitorg/journal/nanosiene ISS: 2381-1234 (Print); ISS: 2381-1242 (Online) Some Properties on ano Topology Indued by Graphs Arafa asef 1 Abd El Fattah
More informationAdvances in Radio Science
Advanes in adio Siene 2003) 1: 99 104 Copernius GmbH 2003 Advanes in adio Siene A hybrid method ombining the FDTD and a time domain boundary-integral equation marhing-on-in-time algorithm A Beker and V
More informationEVALUATION OF RESONANCE PROPERTIES OF THE HUMAN BODY MODELS SITUATED IN THE RF FIELD
EVALUATION OF RESONANCE PROPERTIES OF THE HUMAN BODY MODELS SITUATED IN THE RF FIELD Elena COCHEROVA 1, Gabriel MALICKY 1, Peter KUPEC 1, Vladimir STOFANIK 1, Joze PUCIK 1 1 Institute o Eletronis and Photonis,
More informationScalable system level synthesis for virtually localizable systems
Salable system level synthesis for virtually loalizable systems Nikolai Matni, Yuh-Shyang Wang and James Anderson Abstrat In previous work, we developed the system level approah to ontroller synthesis,
More information