Enumerating Prime Implicants of Propositional Formulae in Conjunctive Normal Form

Size: px
Start display at page:

Download "Enumerating Prime Implicants of Propositional Formulae in Conjunctive Normal Form"

Transcription

1 Enumerating Prime Implicants of Propositional Formulae in Conjunctive Normal Form Said Jabbour 1, Joao Marques-Silva 2, Lakhdar Sais 1, and Yakoub Salhi 1 1 CRIL, Université d Artois & CNRS, Lens, France {jabbour,sais,salhi}@cril.fr 2 CASL, University College Dublin, Ireland jpms@ucd.ie Abstract. In this paper, a new approach for enumerating the set prime implicants (PI) of a Boolean formula in conjunctive normal form (CNF) is proposed. It is based on an encoding of the input formula as a new one whose models correspond to the set of prime implicants of the original theory. This first PI enumeration approach is then enhanced by an original use of the boolean functions or gates usually involved in many CNF instances encoding real-world problems. Experimental evaluation on several classes of CNF instances shows the feasibility of our proposed framework. 1 Introduction The problem of enumerating prime implicants (PIs) of Boolean functions is an important research topic from the early days of computer science. It was used in the context of boolean function minimization by Quine [29, 28] and McCluskey [23]. This first application of the prime implicant canonical form is important as it allows to reduce digital circuit size and cost while improving the computing speed (e.g. [34]). In addition to digital circuit analysis and optimisation, PIs have found several other application domains including fault tree analysis [6, 11], bioinformatics [1], databases [10], model based diagnosis [8], knowledge representation and reasoning [4]. The computation of prime implicants is also important in many subfields of artificial intelligence such us knowledge compilation [3, 7], automated and non-monotonic reasoning [14], multi-agent systems [33]. Unfortunately, the problem of generating all prime implicants of a given propositional theory is a highly complex task. First, the number of prime implicants of a given theory can be exponential in the size of the theory, while finding just one prime implicant is an NP-hard task. Consequently, enumerating all PIs cannot be done in polynomial total time unless P=NP [16]. Despite this computational bottleneck, several techniques have been proposed in the literature. Many of these PI enumeration techniques are based on some adaptation of the well-known search paradigm, namely branch and bound/backtrack search procedures. Additionally, most of these techniques consider propositional formulae in conjunctive normal form (CNF), the standard representation of propositional

2 knowledge bases. One can cite two related adaptations of the well-known DPLL procedure for prime implicants generation [5, 32, 30] or the modification of modern SAT solvers for computing one prime implicant [9]. Concurrently, almost simultaneously, a 0-1 integer linear programming (ILP) formulation [21, 19] was proposed for computing minimum-size prime implicant improving the formulation given in [27]. In addition, a new algorithm for solving ILP was developed, built on top of a CDCL SAT solver. In [26], Palopoli et al. formulated two algorithms for PIs computation. The first one (called Enumerative Prime Implicants/Feasible Solutions - EPI/FS) search for a feasible solution of the linear program and then extract a prime implicant from it. The second variant, called EPI/OS, is obtained by simply adding a suitable minimization function to the ILP formulation. In contrast to PI enumeration based on an adaptation of DPLL-like procedure (e.g. [5, 32]) and ILP based formulations [21, 27], in this paper, we first propose an original approach that rewrite the CNF formula Φ as a new CNF formula Φ such that the prime implicants of Φ correspond the models of the new CNF Φ. In this way, prime implicant enumeration is reduced to the problem of finding all models of a CNF formula. Such correspondence, allows us to benefit from the recent and continuous advances in practical SAT solving at least for finding one prime implicant. From the ILP formulation mentioned above [21, 27], our PI enumeration encoding borrows only the idea of literal renaming. As such renaming substantially increases the number of variables and clauses, in our second contribution, an enhanced encoding is derived thanks to the structural knowledge recovered from the CNF formula. Indeed, by exploiting Boolean functions encoded in the formula, our new encoding allows significant reductions in the number of variables and clauses. Surprisingly, despite the numerous studies on this issue, to the best of our knowledge, there is no available PI enumeration tool. To compare our proposed approach, an additional PI enumeration algorithm is implemented, which is based on adapting previous work (see Section 5). More precisely, the CNF formula Φ is encoded as a partial MaxSAT formula Φ P using a reformulation of the previous mentioned 0-1 ILP model. Thanks to the correspondence between minimal correction subsets (MCSs) of Φ P and the prime implicants of Φ, we exploit the MCSs enumeration tool proposed recently in [22]. The paper is organized as follows. After some preliminary definitions and necessary notations, our SAT-based encoding of PI enumeration problem is provided in Section 3. Then, we describe in Section 4 our structure-based enhancement of PI enumeration. Then, an alternative approach for PIs generation is discussed in Section 5. An extensive experimental evaluation of our proposed approaches is provided (Section 6) before concluding. 2 Preliminary Definitions and Notations We first introduce the satisfiability problem (SAT) and some necessary notations. SAT corresponds to the problem of deciding if a formula of propositional classical

3 logic is consistent or not. It is one of the most studied NP-complete decision problem. We consider the conjunctive normal form (CNF) representation for propositional formulas. A CNF formula Φ is a conjunction of clauses, where a clause is a disjunction of literals. A literal is a positive (p) or negated ( p) propositional variable. The two literals p and p are called complementary. We denote by l the complementary literal of l, i.e., if l = p then l = p and if l = p then l = p. For a set of literals L, L is defined as { l l L}. A CNF formula can also be seen as a set of clauses, and a clause as a set of literals. Let us recall that any propositional formula can be translated to CNF using linear Tseitin s encoding [35]. We denote by V ar(φ) (respectively Lit(Φ)) the set of propositional variables (respectively literals) occurring in Φ. Φ x denotes the formula obtained from Φ by assigning x the truth-value true. Formally Φ x = {c c Φ, {x, x} c = } {c\{ x} c Φ, x c} (that is: the clauses containing x are therefore satisfied and removed; and those containing x are simplified). Φ denotes the formula Φ closed under unit propagation (UP closure), defined recursively as follows: (1) Φ = Φ if Φ does not contain any unit clause, (2) Φ = if Φ contains two unit-clauses {x} and { x}, (3) otherwise, Φ = (Φ x ) where x is the literal appearing in a unit clause of Φ. The set of unit propagated literals by applying UP closure on Φ is denoted UP (Φ). An interpretation B of a propositional formula Φ is a function which associates a value B(p) {0, 1} (0 corresponds to false and 1 to true) to the variables p V ar(φ). An interpretation B of Φ is alternatively represented by a set of literals, i.e., B = x V ar(φ) f(x), where f(x) = x (respectively f(x) = x), if B(x) = true (respectively B(x) = false). A model or an implicant of a formula Φ is an interpretation B that satisfies the formula, noted B Φ. The SAT problem consists in deciding if a given formula admits a model or not. An implicant B of Φ is called a prime implicant (in short PI), iff for all literals l B, B\{l} Φ. We define P I(Φ) as the disjunction of all prime implicant of Φ. Obviously, P I(Φ) is logically equivalent to Φ, while its size might be exponential in the worst case. 3 SAT-based Encoding of PI Enumeration Problem In this section, we describe our SAT-based encoding of the prime implicant enumeration problem. The idea consists in reformulating the PI enumeration problem of a given CNF Φ as the model enumeration problem of a CNF Φ. Our encoding borrows the idea of literals renaming used in the ILP formulations proposed in [27, 19, 20]. Let Φ be a CNF formula. We associate to each element l of Lit(Φ) a propositional variable x l. We define the CNF formula Φ R

4 as the formula obtained from Φ by renaming each literal l in Lit(Φ) by its corresponding propositional variable x l, and by adding the following binary clauses: x p x p (1) p V ar(φ) One can easily see that Φ and Φ R are equisatisfiable. Example 1. Let us consider the following CNF formula: Φ = (p q r) ( p r) (q r). Then, we have Φ R = (x p x q x r ) (x p x r ) (x q x r ) ( x p x p ) ( x q x q ) ( x r x r ). As we can see, the formula Φ R is a conjunction of two monotone CNF formulae. We now propose a new constraint in order to establish a bijection between prime implicants of a CNF formula and the models of the resulting CNF formula. This additional formula M(Φ R ) is defined as follows: M(Φ R ) = x l Cl(Φ R, x l ) (2) l Lit(Φ) where Cl(Φ R, x l ) corresponds to the restriction of Φ R to the clauses containing x l without the latter, i.e., Cl(Φ R, x l ) c Φ R,x l c c \ {x l}. For instance, if we consider again the formula Φ given in Example 1, M(Φ R ) corresponds to the the following formula: (x p (x q x r )) (x p x r ) (x q x r ) (x q (x p x r )) (x r (x p x q )) (x r (x q x p )). Theorem 1. If B is a model of the formula Φ R M(Φ R ), then the set of literals I B = {l Lit(Φ) B(x l ) = 1} is a prime implicant of Φ. Proof. Using the fact that Φ R is nothing else than a renaming of the literals of Φ, we have I B is an implicant of Φ. Assume now that I B is not a prime implicant of Φ. Then, there exists l 0 I such that I B \ {l 0 } is an implicant of Φ. Let B be a Boolean interpretation of Φ R M(Φ R ) defined as follows: { B B(xl ) if l l (x l ) = 0 0 otherwise Clearly, we have B (x l0 ) = B (x l0 ) = 0. Indeed, B (x l0 ) = 0 (see the definition of B ). Let us show that B (x l0 ) = 0. By definition, we have B (x l0 ) = B(x l0 ) because l 0 l 0. As l 0 I B, we have B(x l0 ) = 1. Also, as I B is an implicant of Φ, then it satisfies the binary clause ( x l0 x l0 ) (see formula (1). Consequently, we deduce that B( x l0 ) = 1 i.e. B(x l0 ) = 0. Finally, B (x l0 ) = 0. Using the formula M(Φ R ), B(x l0 Cl(Φ R, x l0 )) = 1 holds. Hence, we have B(Cl(Φ R, x l0 )) = 0, since B(x l0 ) = 1 because l 0 I B. Since B is obtained from B by setting the truth value of x l0 to 0 (B (x l0 ) = 0), then B(Cl(Φ R, x l0 )) = B (Cl(Φ R, x l0 )) = 0. We then obtain that B is not a model of Φ R and we get a contradiction with I B \ {l 0 } is an implicant of Φ. Therefore, I is a prime implicant of Φ.

5 Theorem 2. Let I be a prime implicant of Φ and B a Boolean interpretation of Φ R M(Φ R ) defined as follows: { 1 if l I B(x l ) = 0 otherwise Then, B is a model of Φ R M(Φ R ). Proof. Clearly, B is a model of Φ R because I is a prime implicant of Φ. We now show that B is also a model of M(Φ R ). Let l be a literal in Lit(Φ). If B(x l ) = 0 then B(x l Cl(Φ R, x l )) = 1 holds. Otherwise, we have B(x l ) = 1 and l I. If B(Cl(Φ R, x l )) = 1 then I \ {l} is an implicant of Φ and we get a contradiction. Therefore, B(Cl(Φ R, x l )) = 0 holds. Consequently, B is a model of M(Φ R ) and then of Φ R M(Φ R ). Corollary 1. The number of prime implicants of Φ is equal to the number of models of Φ R M(Φ R ). Proof. For all B 1 and B 2 two different models of Φ R M(Φ R ), I B1 I B2 holds. Thus, using Theorem 1, we obtain that the number of prime implicants of Φ is greater than or equal to the number of models of Φ R M(Φ R ). Moreover, using Theorem 2, the number of prime implicants of Φ is smaller than or equal to the number of models of Φ R M(Φ R ). Therefore, the number of prime implicants of Φ is equal to the number of models of Φ R M(Φ R ). 4 Structure-Based Enhancement of PI Enumeration CNF formulae encoding real words problems usually involve a large fraction of clauses encoding different kind of boolean functions or gates. These Boolean functions result from the problem specification itself or introduced during the CNF transformation using the well-known Tseitin extension principle [35]. Tseitin s encoding consists in introducing fresh variables to represent sub-formulae in order to represent their truth values. For example, given a Boolean formula, containing the variables a and b, and v a fresh variable, one can add the definition v a b (called extension) to the formula while preserving satisfiability. Tseitin s extension principle is at the basis of the linear transformation of general Boolean formulae into CNF. Boolean functions express strong relationships between variables, the goal of this section is to show how such variable dependencies can be exploited in the context of prime impliquant generation. Let us first introduce some formal definitions and notations about Boolean functions or gates. A Boolean function or Gate is an expression of the form l = f(l 1,..., l n ), where l, l 1,..., l n are literals and f is a logical connective among {,, }. It allows us to express that the truth value of l is determined by f(l 1,..., l n ). According to f, a Boolean gate can be defined as a conjunction of clauses as follows:

6 1. l = (l 1,..., l n ) represents the set of clauses { l 1 l n l, l l 1,..., l l n }; 2. l = (l 1,..., l n ) represents the set of clauses { l l 1 l n, l 1 l,..., l n l }; 3. l = (l 1,..., l n ) represents the following equivalence chain (also called biconditional formula) (l l 1,..., l n ). Let us note that the Boolean functions using the connective ( -gates) and those using ( -gates) are dual. Indeed, any Boolean gate l = (l 1,..., l n ) (resp. l = (l 1,..., l n )) is equivalent to l = ( l 1,..., l n ) (resp. l = ( l 1,..., l n )). It is also important to note that for the third Boolean function (equivalence chain), its equivalent representation in CNF leads to a huge number of clauses (2 n clauses). In general, we say that an expression l = f(l 1,..., l n ) is a gate of a CNF formula Φ if it is a logical consequence of Φ. Unfortunately, this deduction problem is Co-NP Complete. However, several recent contributions addressed the issue of recovering Boolean functions from CNF formulae (e.g. [25, 15, 31, 13, 2]). In [25], a polynomial and syntactical approach, which recovers only Boolean functions implicitly present in the CNF formula. Another detection technique is proposed by Roy et al. in [31]. It operates by translating the gate matching problem into one of recognizing sub hypergraph isomorphism. This approach is clearly intractable. In [15], a new technique based on deduction restricted to unit propagation process is proposed. It extends the syntactical approach [25] and allows the detection of some hidden Boolean functions. Given a formula CNF Φ, we can use two different methods for detecting Boolean functions encoded by the clauses of Φ. The first detection method, called syntactical method, is a pattern matching approach that allows us to detect the Boolean functions that appear directly in the structure of the CNF formula [25]. Example 2. Let Φ {(y x 1 x 2 x 3 ), ( y x 1 ), ( y x 2 ), ( y x 3 )}. In this example, we can detect syntactically the function y = (x 1, x 2, x 3 ). The second method is a semantic detection approach where the functions are detected using Unit Propagation (UP) [15]. Indeed, this method allows us to detect hidden Boolean functions linearly in the size of the CNF. Example 3. Let Φ {(y x 1 x 2 x 3 ), ( y x 1 ), ( x 1 x 4 ), ( x 4 x 2 ), ( x 2 x 5 ), ( x 4 x 5 x 3 )}. In this example, UP (Φ y) = {x 1, x 4, x 2, x 5, x 3 } the set of unit propagated (UP) literals and we have the clause c = (y x 1 x 2 x 3 ) Φ which is such that c\{y} UP (Φ y). So, we can discover the Boolean function y = (x 1, x 2, x 3 ), that the above syntactical method does not help us to discover.

7 In our implementation, we exploit the semantic or UP-based approach proposed in [15]. Consequently, the boolean function of the form l = (l 1,..., l n ) is not considered in our experiments. Let us now describe how these Boolean functions can be used to improve the PI enumeration CNF encoding. Proposition 1. Let Φ be a CNF formula and l = f(l 1,..., l n ) a gate of Φ. then, for every prime implicant I, we have either l I or l I. Proof. Assume that there exists a prime implicant I such that both l / I and l / I hold. Consider I an extension of I which is obtained by assigning truth values to all the literals l 1,..., l n and without assigning any truth value to l. Thus, we have either I (f(l 1,..., l n )) = 0 or I (f(l 1,..., l n )) = 1. If I (f(l 1,..., l n )) = 0 then I {l } is a counter-model of Φ and we get a contradiction. Otherwise, I { l } is a counter-model of Φ and we also get contradiction. Let Φ be a CNF formula and {l 1 = e 1,..., l n = e n } a set of its gates. Using Proposition 1, we know that it is not necessary to associate in our encoding fresh propositional variables (of the form x l ) to the literals in S = {l 1,..., l n, l 1,..., l n }. In this case, Φ R is obtained by renaming only the literals in Lit(Φ) \ S and by redefining the formula (1) as follows: We also redefine (2) as follows: M(Φ R ) = p V ar(φ)\s l Lit(Φ)\S x p x p (3) x l Cl(Φ R, x l ) (4) Proposition 2. Let Φ be a CNF formula and l = (l 1,..., l n ) a gate of Φ. then, for every prime implicant I, we have either {l, l 1,..., l n } I or l I. Proof. In the same way as the proof of Proposition 1. Using Proposition 2, one can reduce the formula M(Φ R ). Indeed, given a Boolean gate l = (l 1,..., l n ) of Φ, every formula of the form x l Cl(Φ R, x l ) where l {l 1,..., l n } can be reduced as follows: (l (x l Cl(Φ R, x l, l ))) (5) where Cl(Φ R, x l, l ) corresponds to the formula c Φ R, l / c,x l c c\{x l, l }. This redefinition comes from the fact that if l is in a prime implicant I then l 1,..., l n are also in I. Thus, it is not necessary to reduce a model by removing literals in {l 1, l 1,..., l n, l n } when l is true. In the same way, we have also the following proposition:

8 Proposition 3. Let Φ be a CNF formula and l = (l 1,..., l n ) a gate of Φ. then, for every prime implicant I and every l {l, l 1,..., l n }, we have either l I or l I. Using Proposition 3, we know that it is not necessary to associate in our encoding fresh propositional variables (of the form x l ) to the literals appearing in the equivalence chains. Other related well known XOR constraints of the form (l 1,..., l n ) can be exploited in the same way. Their detection can be done using a pattern matching approach [25]. In this paper, the integration to our encoding of these specific Boolean functions is left as an interesting perspective. 5 Prime Implicant Enumeration: Alternative Approaches As mentioned in the introduction, to the best of our knowledge, there is no available PI enumeration tool. To compare our proposed approach, we discuss an alternative that will be used in our comparative experimental evaluation (Section 6). mcsls One can envision a number of alternative approaches for enumerating prime implicants, by exploiting the 0-1 ILP model for computing a minimum size prime implicant [19]. Given a CNF formula Φ, let Φ i denote the 0-1 ILP model associated with Φ: minimize subject to l Lit(Φ) x l x l 1 for c Φ (6) l c x p + x p 1 for p V ar(φ) (7) This model can be re-formulated as a partial MaxSAT formula [17] Φ P. A simple observation is that any minimal correction subset (MCS) of Φ P is a prime implicant of Φ. Therefore, a tool capable of enumerating the MCSes Φ P can be used for enumerating the prime implicants of Φ. A number of approaches have been proposed in recent years for enumerating MCSes [18, 24, 22]. These either use MaxSAT [18, 24] or dedicated algorithms [22], with recent results indicating that the dedicated algorithms outperform MaxSAT-based solutions. 6 Experiments In this section, we present an experimental evaluation of our approach which consists in enumerating prime implicants of a CNF formula. As described above, our transformation allows to translate the problem of enumerating prime implicants of a formula Φ to that of enumerating models of a new formula Φ. In this context, we use a modified version of MiniSAT [12] solver to enumerate the

9 models of Φ. Each time a model is found a prime implicant is extracted and the clause representing the negation of the model is added to seek for the next model until Φ becomes unsatisfiable. In order to evaluate the performances of our approach, we consider a comparison with the prime implicants enumerator (mcsls described in Section 5. All the experimental results presented in this section have been obtained with a Quad-core Intel Xeon X5550 (2.66GHz, 32 GB RAM) cluster. Our experiments are conducted on benchmarks coming from 2012 MaxSAT Evaluation 3. More precisely, we consider the enumeration of prime implicants of the hard parts of partial MaxSAT instances. 1e+07 1e+08 1e+06 1e e+06 #Vars (Trans.) #Cls (Trans.) e+06 #Vars (Orig.) e+06 1e+07 #Cls (Orig.) Fig. 1. Size of the Encoding With Boolean Gates Let us first illustrate the size of our encoding enhanced with Boolean functions (or gates). In the scatter plot of Figure 1 (left hand side), each dot (x, y) represents the number of variables x of the original (Orig.) formula and the number of variables y of new formula (Trans.) respectively. As we can see, except for some few instances, the resulting encoding is, in general, of reasonable size (in many cases, the number of variables does not exceed ten times the original one). The same observation can be made on the number of clauses (see Figure 1 - right hand side). Figure 2 shows the size of the encoding with and without using Boolean gates. As we can see, exploiting Boolean functions allows us to improve the encoding both in the number of variables and clauses. All the dots are below the diagonal. On a representative sample of instances, Table 6 highlights some characteristics of the original formula and of those obtained using our encodings. In column 1, 2 and 3, we report respectively the name of the instance, its number of variables (#vars) and clauses (#cls). Columns 4 and 5 provide the increasing factor of the number of variables ( #vars) and clauses ( #cls) of the encoding without using Boolean Gates (P I-Encoding). In column 6, we provide the number of Boolean functions or Gates (#bg) detected from the original instance. The two 3 MaxSAT Evaluations:

10 1e+07 1e+08 1e+06 1e+07 #Vars with BG (Trans.) #Cls with BG (Trans.) 1e e+06 1e+07 #Vars without BG (Trans.) e+06 1e+07 1e+08 #Cls without BG (Trans.) Fig. 2. Size of the Encoding With and Without Boolean Gates last columns show the increasing factor of the number of variables and clauses obtained using our enhanced encoding (with Boolean gates - P I-Encoding + BG). We ran the two prime implicant enumerators, primp (our encoding without Boolean gates), mcsls (see Section 5) on the set of the 497 instances taken from the 2012 MaxSAT evaluation. In the first experiment, we limit the size of the output (number of prime implicants) to PIs and we compare the time needed for each approach to output this number of PIs or all PIs of the formula whenever it does not exceed PIs (with a time out of 1 hour). Figure 3, provides a comparison between mcsls and our primp approach. One can see that the majority of the dots are under the diagonal which illustrates the ability of primp to efficiently generate prime implicants compared to mcsls primp mcsls Fig. 3. mcls vs primp: CPU Time Needed to Compute PIs For a deeper analysis, we ran primp and mcsls with a time limit fixed to 1 hour. We compare the number of prime implicants found by each PIs enumerator. In Figure 4, each dot (x, y) represents the number of prime implicants found

11 Original Instance PI-Encoding PI-Encoding+BG instances #vars #cls (#vars) (#cls) #bg (#vars) (#cls) normalized-f20c10b 020 area delay ,89 6, ,86 1,90 normalized-f20c10b 014 area delay ,12 6, ,86 1,97 normalized-fir09 area partials ,11 22, ,33 2,03 normalized-f20c10b 025 area delay ,81 11, ,20 6,52 normalized-fir04 area delay ,07 9, ,81 2,86 normalized-fir01 area opers ,62 7, ,07 3,45 normalized-f20c10b 003 area delay ,88 6, ,80 1,80 normalized-fir05 area partials ,91 28, ,96 11,59 normalized-fir10 area partials ,67 18, ,21 4,63 normalized-m r ,00 903, ,00 903,00 normalized-m r , , , ,00 normalized-max1024.pi ,97 49,08 0 5,97 49,08 normalized-m r , , , ,00 15tree801posib ,46 6, ,46 1,44 15tree201posib ,64 6, ,47 1,44 10tree515p ,63 6, ,42 1,40 15tree901p ,92 7, ,32 1,28 10tree315p ,70 6, ,44 1,42 normalized-s pb ,66 9, ,91 7,01 normalized-s pb ,71 9, ,94 7,04 normalized-hanoi ,94 4, ,81 2,28 normalized-ssa ,51 3, ,14 1,15 normalized-ii32e ,98 74, ,98 74,61 normalized-ii16b ,29 11,30 0 7,29 11,30 normalized-par32-2-c ,56 8, ,92 8,12 normalized-ii32e ,61 89, ,61 89,88 normalized-par32-4-c ,57 8, ,95 8,14 normalized-ii32b ,29 85, ,29 85,54 normalized-ii16b ,72 10,79 0 6,72 10,79 splitedreads 158.matrix ,35 11,22 0 9,35 11,22 splitedreads 0.matrix ,53 10, ,53 10,94 SU3 simp-genos.haps ,75 6, ,75 6,14 simp-ibd ,33 6, ,33 6,58 simp-ibd ,10 6, ,10 6,69 simp-ibd ,38 6, ,32 6,89 simp-test chr10 JPT ,43 6, ,43 6,85 1knt.5pti.g.wcnf.t ,00 1,00 0 2,00 1,00 2knt.5pti.g.wcnf.t ,00 1,00 0 2,00 1,00 6ebx.1era.g.wcnf.t ,00 1,00 0 2,00 1,00 3ebx.1era.g.wcnf.t ,00 1,00 0 2,00 1,00 Table 1. Size of the Encoding: Highlighting Results

12 by mcsls (x) and primp (y). This experiment confirms the efficiency of our approach. Indeed, the enumerator primp outperforms mcsls. Note that primp is able to generate 10 times more prime implicants than mcsls on several instances. 1e+08 1e+07 1e+06 primp e+06 mcsls Fig. 4. mcsls vs primp : Number of Generated PIs in less than 1 hour In our last experiment, we evaluate our encoding enhanced with Boolean gates. Figure 5, compares primp encoding (with and without Boolean gates) in terms of CPU time (in seconds) needed to generate the first PIs, under a time out of 1 hour. Note that the CPU time (seconds) includes the time needed for recovering Boolean gates from the CNF formula. The results demonstrate that exploiting Boolean functions makes significant improvements. Indeed, most of the dots are above the diagonal. Dots that are near the diagonal correspond to instances that involve a marginal number of Boolean functions primp without BG primp with BG Fig. 5. primp with/without Boolean gates: CPU Time Needed to Compute PIs

13 7 Conclusion and Future Works In this paper, we propose a new approach for enumerating the set of prime implicants (PIs) of a Boolean formula in conjunctive normal form (CNF). It reformulates the original PI generation problem as a model generation problem of a new CNF formula. We also propose an essential interesting improvement of the encoding using Boolean functions recovered from the original formula. A comparative evaluation of our approach with an other alternative based PI generation method show significant improvements on several classes of CNF instances. This work open several issues for future research. We first plan to study how other interesting Boolean functions such as equivalence chain and XOR constraints can be recognized and integrated in our framework. Exploiting other structural properties (e.g. monotone literals) might lead to further improvements. To be exhaustive, we also plan to implement several other alternatives for computing prime implicants. References 1. V. Acuña, P. V. Milreu, L. Cottret, A. Marchetti-Spaccamela, L. Stougie, and M.- F. Sagot. Algorithms and complexity of enumerating minimal precursor sets in genome-wide metabolic networks. Bioinformatics, 28(19): , G. Audemard and L. Saïs. Circuit based encoding of CNF formula. In International Conference on Theory and Applications of Satisfiability Testing SAT 2007, pages 16 21, Y. Boufkhad, E. Gregoire, P. Marquis, and L. Sais. Tractable cover compilations. In Proceedings of the Fifteenth International Joint Conference on Artificial Intelligence (IJCAI 97), pages , M. Cadoli and F. M. Donini. A survey on knowledge compilation. AI Commun., 10(3-4): , T. Castell. Computation of prime implicates and prime implicants by a variant of the Davis and Putnam procedure. In ICTAI, pages , O. Coudert and J. Madre. Fault tree analysis: prime implicants and beyond. In Reliability and Maintainability Symposium, pages , Jan A. Darwiche and P. Marquis. A knowledge compilation map. J. Artif. Intell. Res. (JAIR), 17: , J. de Kleer, A. K. Mackworth, and R. Reiter. Characterizing diagnoses. In Proceedings of the 8th National Conference on Artificial Intelligence (AAAI 90), pages , D. Déharbe, P. Fontaine, D. L. Berre, and B. Mazure. Computing prime implicants. In FMCAD, pages 46 52, A. del Val. Tractable databases: How to make propositional unit resolution complete through compilation. In Proceedings of the 4th International Conference on Principles of Knowledge Representation and Reasoning (KR 94). Bonn, Germany, May 24-27, 1994, pages , Y. Dutuit and A. Rauzy. Exact and truncated computations of prime implicants of coherent and non-coherent fault trees within Aralia. Reliability Engineering and System Safety, 58(2): , N. Eén and N. Sörensson. An extensible SAT-solver. In SAT, pages , 2003.

14 13. Z. Fu and S. Malik. Extracting logic circuit structure from conjunctive normal form descriptions. In 20th International Conference on VLSI Design (VLSI Design 2007), Sixth International Conference on Embedded Systems (ICES 2007), 6-10 January 2007, Bangalore, India, pages 37 42, M. Ginsberg. A circumscriptive theorem prover. In M. Reinfrank, J. Kleer, M. Ginsberg, and E. Sandewall, editors, Non-Monotonic Reasoning, volume 346 of Lecture Notes in Computer Science, pages Springer Berlin Heidelberg, É. Grégoire, R. Ostrowski, B. Mazure, and L. Saïs. Automatic extraction of functional dependencies. In Proceedings of the Seventh International Conference on Theory and Applications of Satisfiability Testing (SAT 04), pages , V. Gurvich and L. Khachiyan. On generating the irredundant conjunctive and disjunctive normal forms of monotone boolean functions. Discrete Applied Mathematics, 96-97(1): , Oct F. Heras, J. Larrosa, S. de Givry, and T. Schiex and 2007 Max-SAT evaluations: Contributed instances. JSAT, 4(2-4): , M. H. Liffiton and K. A. Sakallah. Algorithms for computing minimal unsatisfiable subsets of constraints. J. Autom. Reasoning, 40(1):1 33, V. M. Manquinho, P. Flores, J. Marques-Silva, and A. L. Oliveira. Prime implicant computation using satisfiability algorithms. In in Proc. of the IEEE International Conference on Tools with Artificial Intelligence, pages , V. M. Manquinho, A. L. Oliveira, and J. Marques-Silva. Models and algorithms for computing minimum-size prime implicants. In In Proc. International Workshop on Boolean Problems (IWBP 98, J. Marques-Silva. On computing minimum size prime implicants. In International Workshop on Logic Synthesis, J. Marques-Silva, F. Heras, M. Janota, A. Previti, and A. Belov. On computing minimal correction subsets. In IJCAI, J. McCluskey, E.J. Minimization of boolean functions. Bell System Technical Journal, 35(6): , A. Morgado, M. H. Liffiton, and J. Marques-Silva. MaxSAT-based MCS enumeration. In HVC, pages , R. Ostrowski, É. Grégoire, B. Mazure, and L. Saïs. Recovering and exploiting structural knowledge from cnf formulas. In Proceedings of the Eighth International Conference on Principles and Practice of Constraint Programming(CP 02), pages , Ithaca (N.Y.), L. Palopoli, F. Pirri, and C. Pizzuti. Algorithms for selective enumeration of prime implicants. Artificial Intelligence, 111(12):41 72, C. Pizzuti. Computing prime implicants by integer programming. In Proceedings of the 8th International Conference on Tools with Artificial Intelligence, ICTAI, pages , Washington, DC, USA, IEEE Computer Society. 28. W. Quine. On cores and prime implicants of truth functions. American Mathematical Monthly, pages , W. V. Quine. The problem of simplifying truth functions. The American Mathematical Monthly, 59(8): , K. Ravi and F. Somenzi. Minimal assignments for bounded model checking. In TACAS, pages 31 45, J. A. Roy, I. L. Markov, and V. Bertacco. Restoring circuit structure from SAT instances. In IWLS, pages , Temecula Creek CA, June 2004.

15 32. R. Schrag. Compilation for critically constrained knowledge bases. In Proceedings of the Thirteenth National Conference on Artificial Intelligence and Eighth Innovative Applications of Artificial Intelligence Conference (AAAI 96), pages , M. Slavkovik and T. Agotnes. A judgment set similarity measure based on prime implicants. In Proceedings of the 13th International Conference on Autonomous Agents and Multiagent Systems, AAMAS 14, page to appear, P. Tison. Generalized consensus theory and applications to the minimization of boolean circuits. IEEE Transactions on Computers, 16(4): , G. Tseitin. On the complexity of derivations in the propositional calculus. In H. Slesenko, editor, Structures in Constructives Mathematics and Mathematical Logic, Part II, pages , 1968.

A Method for Generating All the Prime Implicants of Binary CNF Formulas

A Method for Generating All the Prime Implicants of Binary CNF Formulas A Method for Generating All the Prime Implicants of Binary CNF Formulas Yakoub Salhi CRIL-CNRS, Université d Artois, F-62307 Lens Cedex, France salhi@cril.fr Abstract In this paper, we propose a method

More information

Solvers for the Problem of Boolean Satisfiability (SAT) Will Klieber Aug 31, 2011

Solvers for the Problem of Boolean Satisfiability (SAT) Will Klieber Aug 31, 2011 Solvers for the Problem of Boolean Satisfiability (SAT) Will Klieber 15-414 Aug 31, 2011 Why study SAT solvers? Many problems reduce to SAT. Formal verification CAD, VLSI Optimization AI, planning, automated

More information

A brief introduction to Logic. (slides from

A brief introduction to Logic. (slides from A brief introduction to Logic (slides from http://www.decision-procedures.org/) 1 A Brief Introduction to Logic - Outline Propositional Logic :Syntax Propositional Logic :Semantics Satisfiability and validity

More information

Using Graph-Based Representation to Derive Hard SAT instances

Using Graph-Based Representation to Derive Hard SAT instances Using Graph-Based Representation to Derive Hard SAT instances Gilles Audemard, Saïd Jabbour, and Lakhdar Saïs CRIL - Université d Artois - Lens, France {audemard,jabbour,sais}@cril.fr Abstract. In this

More information

Encoding formulas with partially constrained weights in a possibilistic-like many-sorted propositional logic

Encoding formulas with partially constrained weights in a possibilistic-like many-sorted propositional logic Encoding formulas with partially constrained weights in a possibilistic-like many-sorted propositional logic Salem Benferhat CRIL-CNRS, Université d Artois rue Jean Souvraz 62307 Lens Cedex France benferhat@criluniv-artoisfr

More information

Solving (Weighted) Partial MaxSAT with ILP

Solving (Weighted) Partial MaxSAT with ILP Solving (Weighted) Partial MaxSAT with ILP Carlos Ansótegui 1 and Joel Gabàs 1 DIEI, Univ. de Lleida carlos@diei.udl.cat joel.gabas@diei.udl.cat Abstract. Several combinatorial optimization problems can

More information

On Unit-Refutation Complete Formulae with Existentially Quantified Variables

On Unit-Refutation Complete Formulae with Existentially Quantified Variables On Unit-Refutation Complete Formulae with Existentially Quantified Variables Lucas Bordeaux 1 Mikoláš Janota 2 Joao Marques-Silva 3 Pierre Marquis 4 1 Microsoft Research, Cambridge 2 INESC-ID, Lisboa 3

More information

arxiv: v1 [cs.ai] 8 Jun 2015

arxiv: v1 [cs.ai] 8 Jun 2015 On SAT Models Enumeration in Itemset Mining Said Jabbour and Lakhdar Sais and Yakoub Salhi CRIL-CNRS Univ. Artois F-6237, Lens France {jabbour, sais, salhi}@univ-orleans.fr arxiv:156.2561v1 [cs.ai] 8 Jun

More information

Boosting MCSes Enumeration

Boosting MCSes Enumeration Boosting MCSes Enumeration Éric Grégoire, Yacine Izza, Jean-Marie Lagniez CRIL - Université d Artois & CNRS, Lens, France {gregoire,izza,lagniez}@cril.fr Abstract The enumeration of all Maximal Satisfiable

More information

A Pigeon-Hole Based Encoding of Cardinality Constraints

A Pigeon-Hole Based Encoding of Cardinality Constraints TPLP 13 (4-5): Online Supplement, July 2013. c 2013 [SAID JABBOUR, LAKHDAR SAIS and YAKOUB SALHI] 1 A Pigeon-Hole Based Encoding of Cardinality Constraints Said Jabbour and Lakhdar Sais and Yakoub Salhi

More information

Comp487/587 - Boolean Formulas

Comp487/587 - Boolean Formulas Comp487/587 - Boolean Formulas 1 Logic and SAT 1.1 What is a Boolean Formula Logic is a way through which we can analyze and reason about simple or complicated events. In particular, we are interested

More information

Decision Procedures for Satisfiability and Validity in Propositional Logic

Decision Procedures for Satisfiability and Validity in Propositional Logic Decision Procedures for Satisfiability and Validity in Propositional Logic Meghdad Ghari Institute for Research in Fundamental Sciences (IPM) School of Mathematics-Isfahan Branch Logic Group http://math.ipm.ac.ir/isfahan/logic-group.htm

More information

LOGIC PROPOSITIONAL REASONING

LOGIC PROPOSITIONAL REASONING LOGIC PROPOSITIONAL REASONING WS 2017/2018 (342.208) Armin Biere Martina Seidl biere@jku.at martina.seidl@jku.at Institute for Formal Models and Verification Johannes Kepler Universität Linz Version 2018.1

More information

SAT Solvers: Theory and Practice

SAT Solvers: Theory and Practice Summer School on Verification Technology, Systems & Applications, September 17, 2008 p. 1/98 SAT Solvers: Theory and Practice Clark Barrett barrett@cs.nyu.edu New York University Summer School on Verification

More information

From SAT To SMT: Part 1. Vijay Ganesh MIT

From SAT To SMT: Part 1. Vijay Ganesh MIT From SAT To SMT: Part 1 Vijay Ganesh MIT Software Engineering & SMT Solvers An Indispensable Tactic for Any Strategy Formal Methods Program Analysis SE Goal: Reliable/Secure Software Automatic Testing

More information

Formal Verification Methods 1: Propositional Logic

Formal Verification Methods 1: Propositional Logic Formal Verification Methods 1: Propositional Logic John Harrison Intel Corporation Course overview Propositional logic A resurgence of interest Logic and circuits Normal forms The Davis-Putnam procedure

More information

Propositional Logic: Evaluating the Formulas

Propositional Logic: Evaluating the Formulas Institute for Formal Models and Verification Johannes Kepler University Linz VL Logik (LVA-Nr. 342208) Winter Semester 2015/2016 Propositional Logic: Evaluating the Formulas Version 2015.2 Armin Biere

More information

CS156: The Calculus of Computation

CS156: The Calculus of Computation CS156: The Calculus of Computation Zohar Manna Winter 2010 It is reasonable to hope that the relationship between computation and mathematical logic will be as fruitful in the next century as that between

More information

SAT-Based Formula Simplification

SAT-Based Formula Simplification SAT-Based Formula Simplification Alexey Ignatiev 1, Alessandro Previti 2, and Joao Marques-Silva 1,2 1 INESC-ID, IST, University of Lisbon, Portugal 2 CASL, University College Dublin, Ireland Abstract.

More information

Knowledge base (KB) = set of sentences in a formal language Declarative approach to building an agent (or other system):

Knowledge base (KB) = set of sentences in a formal language Declarative approach to building an agent (or other system): Logic Knowledge-based agents Inference engine Knowledge base Domain-independent algorithms Domain-specific content Knowledge base (KB) = set of sentences in a formal language Declarative approach to building

More information

Lecture 2 Propositional Logic & SAT

Lecture 2 Propositional Logic & SAT CS 5110/6110 Rigorous System Design Spring 2017 Jan-17 Lecture 2 Propositional Logic & SAT Zvonimir Rakamarić University of Utah Announcements Homework 1 will be posted soon Propositional logic: Chapter

More information

Prime Compilation of Non-Clausal Formulae

Prime Compilation of Non-Clausal Formulae Proceedings of the Twenty-Fourth International Joint Conference on Artificial Intelligence (IJCAI 2015) Prime Compilation of Non-Clausal Formulae A. Previti UCD CASL Dublin, Ireland alessandro.previti@ucdconnect.ie

More information

CSE507. Introduction. Computer-Aided Reasoning for Software. Emina Torlak courses.cs.washington.edu/courses/cse507/17wi/

CSE507. Introduction. Computer-Aided Reasoning for Software. Emina Torlak courses.cs.washington.edu/courses/cse507/17wi/ Computer-Aided Reasoning for Software CSE507 courses.cs.washington.edu/courses/cse507/17wi/ Introduction Emina Torlak emina@cs.washington.edu Today What is this course about? Course logistics Review of

More information

VLSI CAD: Lecture 4.1. Logic to Layout. Computational Boolean Algebra Representations: Satisfiability (SAT), Part 1

VLSI CAD: Lecture 4.1. Logic to Layout. Computational Boolean Algebra Representations: Satisfiability (SAT), Part 1 VLSI CAD: Logic to Layout Rob A. Rutenbar University of Illinois Lecture 4.1 Computational Boolean Algebra Representations: Satisfiability (SAT), Part 1 Some Terminology Satisfiability (called SAT for

More information

COMP219: Artificial Intelligence. Lecture 20: Propositional Reasoning

COMP219: Artificial Intelligence. Lecture 20: Propositional Reasoning COMP219: Artificial Intelligence Lecture 20: Propositional Reasoning 1 Overview Last time Logic for KR in general; Propositional Logic; Natural Deduction Today Entailment, satisfiability and validity Normal

More information

Introduction to Artificial Intelligence Propositional Logic & SAT Solving. UIUC CS 440 / ECE 448 Professor: Eyal Amir Spring Semester 2010

Introduction to Artificial Intelligence Propositional Logic & SAT Solving. UIUC CS 440 / ECE 448 Professor: Eyal Amir Spring Semester 2010 Introduction to Artificial Intelligence Propositional Logic & SAT Solving UIUC CS 440 / ECE 448 Professor: Eyal Amir Spring Semester 2010 Today Representation in Propositional Logic Semantics & Deduction

More information

Automated Program Verification and Testing 15414/15614 Fall 2016 Lecture 2: Propositional Logic

Automated Program Verification and Testing 15414/15614 Fall 2016 Lecture 2: Propositional Logic Automated Program Verification and Testing 15414/15614 Fall 2016 Lecture 2: Propositional Logic Matt Fredrikson mfredrik@cs.cmu.edu October 17, 2016 Matt Fredrikson Propositional Logic 1 / 33 Propositional

More information

A Unit Resolution Approach to Knowledge Compilation. 2. Preliminaries

A Unit Resolution Approach to Knowledge Compilation. 2. Preliminaries A Unit Resolution Approach to Knowledge Compilation Arindama Singh and Manoj K Raut Department of Mathematics Indian Institute of Technology Chennai-600036, India Abstract : Knowledge compilation deals

More information

Critical Reading of Optimization Methods for Logical Inference [1]

Critical Reading of Optimization Methods for Logical Inference [1] Critical Reading of Optimization Methods for Logical Inference [1] Undergraduate Research Internship Department of Management Sciences Fall 2007 Supervisor: Dr. Miguel Anjos UNIVERSITY OF WATERLOO Rajesh

More information

Propositional and First Order Reasoning

Propositional and First Order Reasoning Propositional and First Order Reasoning Terminology Propositional variable: boolean variable (p) Literal: propositional variable or its negation p p Clause: disjunction of literals q \/ p \/ r given by

More information

On Computing Maximal Subsets of Clauses that Must Be Satisfiable with Possibly Mutually-Contradictory Assumptive Contexts

On Computing Maximal Subsets of Clauses that Must Be Satisfiable with Possibly Mutually-Contradictory Assumptive Contexts Proceedings of the Twenty-Ninth AAAI Conference on Artificial Intelligence On Computing Maximal Subsets of Clauses that Must Be Satisfiable with Possibly Mutually-Contradictory Assumptive Contexts Philippe

More information

Interleaved Alldifferent Constraints: CSP vs. SAT Approaches

Interleaved Alldifferent Constraints: CSP vs. SAT Approaches Interleaved Alldifferent Constraints: CSP vs. SAT Approaches Frédéric Lardeux 3, Eric Monfroy 1,2, and Frédéric Saubion 3 1 Universidad Técnica Federico Santa María, Valparaíso, Chile 2 LINA, Université

More information

A Logically Complete Reasoning Maintenance System Based on a Logical Constraint Solver

A Logically Complete Reasoning Maintenance System Based on a Logical Constraint Solver A Logically Complete Reasoning Maintenance System Based on a Logical Constraint Solver J.C. Madre and O. Coudert Bull Corporate Research Center Rue Jean Jaurès 78340 Les Clayes-sous-bois FRANCE Abstract

More information

Foundations of Artificial Intelligence

Foundations of Artificial Intelligence Foundations of Artificial Intelligence 31. Propositional Logic: DPLL Algorithm Malte Helmert and Gabriele Röger University of Basel April 24, 2017 Propositional Logic: Overview Chapter overview: propositional

More information

Tecniche di Verifica. Introduction to Propositional Logic

Tecniche di Verifica. Introduction to Propositional Logic Tecniche di Verifica Introduction to Propositional Logic 1 Logic A formal logic is defined by its syntax and semantics. Syntax An alphabet is a set of symbols. A finite sequence of these symbols is called

More information

The Calculus of Computation: Decision Procedures with Applications to Verification. Part I: FOUNDATIONS. by Aaron Bradley Zohar Manna

The Calculus of Computation: Decision Procedures with Applications to Verification. Part I: FOUNDATIONS. by Aaron Bradley Zohar Manna The Calculus of Computation: Decision Procedures with Applications to Verification Part I: FOUNDATIONS by Aaron Bradley Zohar Manna 1. Propositional Logic(PL) Springer 2007 1-1 1-2 Propositional Logic(PL)

More information

Price: $25 (incl. T-Shirt, morning tea and lunch) Visit:

Price: $25 (incl. T-Shirt, morning tea and lunch) Visit: Three days of interesting talks & workshops from industry experts across Australia Explore new computing topics Network with students & employers in Brisbane Price: $25 (incl. T-Shirt, morning tea and

More information

USING SAT FOR COMBINATIONAL IMPLEMENTATION CHECKING. Liudmila Cheremisinova, Dmitry Novikov

USING SAT FOR COMBINATIONAL IMPLEMENTATION CHECKING. Liudmila Cheremisinova, Dmitry Novikov International Book Series "Information Science and Computing" 203 USING SAT FOR COMBINATIONAL IMPLEMENTATION CHECKING Liudmila Cheremisinova, Dmitry Novikov Abstract. The problem of checking whether a

More information

A Resolution Method for Modal Logic S5

A Resolution Method for Modal Logic S5 EPiC Series in Computer Science Volume 36, 2015, Pages 252 262 GCAI 2015. Global Conference on Artificial Intelligence A Resolution Method for Modal Logic S5 Yakoub Salhi and Michael Sioutis Université

More information

A Knowledge Compilation Map

A Knowledge Compilation Map Journal of Artificial Intelligence Research 17 (2002) 229-264 Submitted 12/01; published 9/02 A Knowledge Compilation Map Adnan Darwiche Computer Science Department University of California, Los Angeles

More information

Basing Decisions on Sentences in Decision Diagrams

Basing Decisions on Sentences in Decision Diagrams Proceedings of the Twenty-Sixth AAAI Conference on Artificial Intelligence Basing Decisions on Sentences in Decision Diagrams Yexiang Xue Department of Computer Science Cornell University yexiang@cs.cornell.edu

More information

Propositional Fragments for Knowledge Compilation and Quantified Boolean Formulae

Propositional Fragments for Knowledge Compilation and Quantified Boolean Formulae Propositional Fragments for Knowledge Compilation and Quantified Boolean Formulae Sylvie Coste-Marquis and Daniel Le Berre and Florian Letombe and Pierre Marquis CRIL/CNRS, Université d Artois, F-62307

More information

Cardinality Networks: a Theoretical and Empirical Study

Cardinality Networks: a Theoretical and Empirical Study Constraints manuscript No. (will be inserted by the editor) Cardinality Networks: a Theoretical and Empirical Study Roberto Asín, Robert Nieuwenhuis, Albert Oliveras, Enric Rodríguez-Carbonell Received:

More information

Semantically Guided Theorem Proving for Diagnosis Applications

Semantically Guided Theorem Proving for Diagnosis Applications Semantically Guided Theorem Proving for Diagnosis Applications Peter Baumgartner Peter Fröhlich Univ. Koblenz Universität Hannover Inst. f. Informatik Abstract In this paper we demonstrate how general

More information

Clause/Term Resolution and Learning in the Evaluation of Quantified Boolean Formulas

Clause/Term Resolution and Learning in the Evaluation of Quantified Boolean Formulas Journal of Artificial Intelligence Research 1 (1993) 1-15 Submitted 6/91; published 9/91 Clause/Term Resolution and Learning in the Evaluation of Quantified Boolean Formulas Enrico Giunchiglia Massimo

More information

Propositional Logic. Methods & Tools for Software Engineering (MTSE) Fall Prof. Arie Gurfinkel

Propositional Logic. Methods & Tools for Software Engineering (MTSE) Fall Prof. Arie Gurfinkel Propositional Logic Methods & Tools for Software Engineering (MTSE) Fall 2017 Prof. Arie Gurfinkel References Chpater 1 of Logic for Computer Scientists http://www.springerlink.com/content/978-0-8176-4762-9/

More information

Compiling Stratified Belief Bases

Compiling Stratified Belief Bases Compiling Stratified Belief Bases Sylvie Coste-Marquis 1 and Pierre Marquis 2 Abstract. Many coherence-based approaches to inconsistency handling within propositional belief bases have been proposed so

More information

Computational Logic. Davide Martinenghi. Spring Free University of Bozen-Bolzano. Computational Logic Davide Martinenghi (1/30)

Computational Logic. Davide Martinenghi. Spring Free University of Bozen-Bolzano. Computational Logic Davide Martinenghi (1/30) Computational Logic Davide Martinenghi Free University of Bozen-Bolzano Spring 2010 Computational Logic Davide Martinenghi (1/30) Propositional Logic - sequent calculus To overcome the problems of natural

More information

Title: Logical Agents AIMA: Chapter 7 (Sections 7.4 and 7.5)

Title: Logical Agents AIMA: Chapter 7 (Sections 7.4 and 7.5) B.Y. Choueiry 1 Instructor s notes #12 Title: Logical Agents AIMA: Chapter 7 (Sections 7.4 and 7.5) Introduction to Artificial Intelligence CSCE 476-876, Fall 2018 URL: www.cse.unl.edu/ choueiry/f18-476-876

More information

Chapter 7 R&N ICS 271 Fall 2017 Kalev Kask

Chapter 7 R&N ICS 271 Fall 2017 Kalev Kask Set 6: Knowledge Representation: The Propositional Calculus Chapter 7 R&N ICS 271 Fall 2017 Kalev Kask Outline Representing knowledge using logic Agent that reason logically A knowledge based agent Representing

More information

An instance of SAT is defined as (X, S)

An instance of SAT is defined as (X, S) SAT: Propositional Satisfiability 22c:45 Artificial Intelligence Russell & Norvig, Ch. 7.6 Validity vs. Satisfiability Validity: A sentence is valid if it is true in every interpretation (every interpretation

More information

A Study on Monotone Self-Dual Boolean Functions

A Study on Monotone Self-Dual Boolean Functions A Study on Monotone Self-Dual Boolean Functions Mustafa Altun a and Marc D Riedel b a Electronics and Communication Engineering, Istanbul Technical University, Istanbul, Turkey 34469 b Electrical and Computer

More information

Agenda. Artificial Intelligence. Reasoning in the Wumpus World. The Wumpus World

Agenda. Artificial Intelligence. Reasoning in the Wumpus World. The Wumpus World Agenda Artificial Intelligence 10. Propositional Reasoning, Part I: Principles How to Think About What is True or False 1 Introduction Álvaro Torralba Wolfgang Wahlster 2 Propositional Logic 3 Resolution

More information

Artificial Intelligence Chapter 7: Logical Agents

Artificial Intelligence Chapter 7: Logical Agents Artificial Intelligence Chapter 7: Logical Agents Michael Scherger Department of Computer Science Kent State University February 20, 2006 AI: Chapter 7: Logical Agents 1 Contents Knowledge Based Agents

More information

CS156: The Calculus of Computation Zohar Manna Autumn 2008

CS156: The Calculus of Computation Zohar Manna Autumn 2008 Page 3 of 52 Page 4 of 52 CS156: The Calculus of Computation Zohar Manna Autumn 2008 Lecturer: Zohar Manna (manna@cs.stanford.edu) Office Hours: MW 12:30-1:00 at Gates 481 TAs: Boyu Wang (wangboyu@stanford.edu)

More information

A New Look at BDDs for Pseudo-Boolean Constraints

A New Look at BDDs for Pseudo-Boolean Constraints Journal of Artificial Intelligence Research 45 (22) 443-48 Submitted 3/2; published /2 A New Look at BDDs for Pseudo-Boolean Constraints Ignasi Abío Robert Nieuwenhuis Albert Oliveras Enric Rodríguez-Carbonell

More information

Detecting Backdoor Sets with Respect to Horn and Binary Clauses

Detecting Backdoor Sets with Respect to Horn and Binary Clauses Detecting Backdoor Sets with Respect to Horn and Binary Clauses Naomi Nishimura 1,, Prabhakar Ragde 1,, and Stefan Szeider 2, 1 School of Computer Science, University of Waterloo, Waterloo, Ontario, N2L

More information

Yet Another Proof of the Strong Equivalence Between Propositional Theories and Logic Programs

Yet Another Proof of the Strong Equivalence Between Propositional Theories and Logic Programs Yet Another Proof of the Strong Equivalence Between Propositional Theories and Logic Programs Joohyung Lee and Ravi Palla School of Computing and Informatics Arizona State University, Tempe, AZ, USA {joolee,

More information

Tutorial 1: Modern SMT Solvers and Verification

Tutorial 1: Modern SMT Solvers and Verification University of Illinois at Urbana-Champaign Tutorial 1: Modern SMT Solvers and Verification Sayan Mitra Electrical & Computer Engineering Coordinated Science Laboratory University of Illinois at Urbana

More information

On the Counting Complexity of Propositional Circumscription

On the Counting Complexity of Propositional Circumscription On the Counting Complexity of Propositional Circumscription Arnaud Durand Miki Hermann Abstract Propositional circumscription, asking for the minimal models of a Boolean formula, is an important problem

More information

Duality in Knowledge Compilation Techniques

Duality in Knowledge Compilation Techniques Duality in Knowledge Compilation Techniques Neil V. Murray 1 and Erik Rosenthal 2 1 Department of Computer Science, State University of New York, Albany, NY 12222, USA, nvm@cs.albany.edu 2 Department of

More information

An Introduction to SAT Solving

An Introduction to SAT Solving An Introduction to SAT Solving Applied Logic for Computer Science UWO December 3, 2017 Applied Logic for Computer Science An Introduction to SAT Solving UWO December 3, 2017 1 / 46 Plan 1 The Boolean satisfiability

More information

Horn upper bounds of random 3-CNF: a computational study

Horn upper bounds of random 3-CNF: a computational study Horn upper bounds of random 3-CNF: a computational study Marina Langlois Robert H. Sloan György Turán 1 Email: mirodo1 sloan gyt@uic.edu University of Illinois at Chicago 1 Also University of Szeged, Hunagary

More information

On the Complexity of Enumerating Pseudo-intents

On the Complexity of Enumerating Pseudo-intents On the Complexity of Enumerating Pseudo-intents Felix Distel a, Barış Sertkaya,b a Theoretical Computer Science, TU Dresden Nöthnitzer Str. 46 01187 Dresden, Germany b SAP Research Center Dresden Chemtnitzer

More information

Unifying hierarchies of fast SAT decision and knowledge compilation

Unifying hierarchies of fast SAT decision and knowledge compilation Unifying hierarchies of fast SAT decision and knowledge compilation Matthew Gwynne and Oliver Kullmann Swansea University, United Kingdom http://cs.swan.ac.uk/~csmg Špindlerův Mlýn, January 28, 2013, SOFSEM

More information

Part 1: Propositional Logic

Part 1: Propositional Logic Part 1: Propositional Logic Literature (also for first-order logic) Schöning: Logik für Informatiker, Spektrum Fitting: First-Order Logic and Automated Theorem Proving, Springer 1 Last time 1.1 Syntax

More information

Towards an Optimal CNF Encoding of Boolean Cardinality Constraints

Towards an Optimal CNF Encoding of Boolean Cardinality Constraints Towards an Optimal CNF Encoding of Boolean Cardinality Constraints (Technical Report) Carsten Sinz Symbolic Computation Group, WSI for Computer Science University of Tübingen, 72076 Tübingen, Germany sinz@informatik.uni-tuebingen.de

More information

Worst-Case Upper Bound for (1, 2)-QSAT

Worst-Case Upper Bound for (1, 2)-QSAT Worst-Case Upper Bound for (1, 2)-QSAT Minghao Yin Department of Computer, Northeast Normal University, Changchun, China, 130117 ymh@nenu.edu.cn Abstract. The rigorous theoretical analysis of the algorithm

More information

Representing Policies for Quantified Boolean Formulas

Representing Policies for Quantified Boolean Formulas Representing Policies for Quantified Boolean Formulas Sylvie Coste-Marquis CRIL Univ. d Artois, Lens, France Hélène Fargier and Jérôme Lang IRIT Univ. Paul Sabatier, Toulouse, France Daniel Le Berre and

More information

Propositional Logic: Methods of Proof (Part II)

Propositional Logic: Methods of Proof (Part II) Propositional Logic: Methods of Proof (Part II) This lecture topic: Propositional Logic (two lectures) Chapter 7.1-7.4 (previous lecture, Part I) Chapter 7.5 (this lecture, Part II) (optional: 7.6-7.8)

More information

EE562 ARTIFICIAL INTELLIGENCE FOR ENGINEERS

EE562 ARTIFICIAL INTELLIGENCE FOR ENGINEERS EE562 ARTIFICIAL INTELLIGENCE FOR ENGINEERS Lecture 10, 5/9/2005 University of Washington, Department of Electrical Engineering Spring 2005 Instructor: Professor Jeff A. Bilmes Logical Agents Chapter 7

More information

Pythagorean Triples and SAT Solving

Pythagorean Triples and SAT Solving Pythagorean Triples and SAT Solving Moti Ben-Ari Department of Science Teaching Weizmann Institute of Science http://www.weizmann.ac.il/sci-tea/benari/ c 2017-18 by Moti Ben-Ari. This work is licensed

More information

Algorithms for Satisfiability beyond Resolution.

Algorithms for Satisfiability beyond Resolution. Algorithms for Satisfiability beyond Resolution. Maria Luisa Bonet UPC, Barcelona, Spain Oaxaca, August, 2018 Co-Authors: Sam Buss, Alexey Ignatiev, Joao Marques-Silva, Antonio Morgado. Motivation. Satisfiability

More information

1 Algebraic Methods. 1.1 Gröbner Bases Applied to SAT

1 Algebraic Methods. 1.1 Gröbner Bases Applied to SAT 1 Algebraic Methods In an algebraic system Boolean constraints are expressed as a system of algebraic equations or inequalities which has a solution if and only if the constraints are satisfiable. Equations

More information

Reasoning with Quantified Boolean Formulas

Reasoning with Quantified Boolean Formulas Reasoning with Quantified Boolean Formulas Martina Seidl Institute for Formal Models and Verification Johannes Kepler University Linz 1 What are QBF? Quantified Boolean formulas (QBF) are formulas of propositional

More information

Proof Complexity of Quantified Boolean Formulas

Proof Complexity of Quantified Boolean Formulas Proof Complexity of Quantified Boolean Formulas Olaf Beyersdorff School of Computing, University of Leeds Olaf Beyersdorff Proof Complexity of Quantified Boolean Formulas 1 / 39 Proof complexity (in one

More information

Improvements to Core-Guided Binary Search for MaxSAT

Improvements to Core-Guided Binary Search for MaxSAT Improvements to Core-Guided Binary Search for MaxSAT A.Morgado F.Heras J.Marques-Silva CASL/CSI, University College Dublin Dublin, Ireland SAT, June 2012 Outline Motivation Previous Algorithms - Overview

More information

Lecture Notes on SAT Solvers & DPLL

Lecture Notes on SAT Solvers & DPLL 15-414: Bug Catching: Automated Program Verification Lecture Notes on SAT Solvers & DPLL Matt Fredrikson André Platzer Carnegie Mellon University Lecture 10 1 Introduction In this lecture we will switch

More information

Preprocessing QBF: Failed Literals and Quantified Blocked Clause Elimination

Preprocessing QBF: Failed Literals and Quantified Blocked Clause Elimination Preprocessing QBF: Failed Literals and Quantified Blocked Clause Elimination Florian Lonsing (joint work with Armin Biere and Martina Seidl) Institute for Formal Models and Verification (FMV) Johannes

More information

On Computing Backbones of Propositional Theories

On Computing Backbones of Propositional Theories On Computing Backbones of Propositional Theories Joao Marques-Silva 1 Mikoláš Janota 2 Inês Lynce 3 1 CASL/CSI, University College Dublin, Ireland 2 INESC-ID, Lisbon, Portugal 3 INESC-ID/IST, Lisbon, Portugal

More information

Logical Agents (I) Instructor: Tsung-Che Chiang

Logical Agents (I) Instructor: Tsung-Che Chiang Logical Agents (I) Instructor: Tsung-Che Chiang tcchiang@ieee.org Department of Computer Science and Information Engineering National Taiwan Normal University Artificial Intelligence, Spring, 2010 編譯有誤

More information

Propositional Logic: Methods of Proof. Chapter 7, Part II

Propositional Logic: Methods of Proof. Chapter 7, Part II Propositional Logic: Methods of Proof Chapter 7, Part II Inference in Formal Symbol Systems: Ontology, Representation, ti Inference Formal Symbol Systems Symbols correspond to things/ideas in the world

More information

The Incremental Satisfiability Problem for a Two Conjunctive Normal Form

The Incremental Satisfiability Problem for a Two Conjunctive Normal Form The Incremental Satisfiability Problem for a Two Conjunctive Normal Form Guillermo De Ita 1, J. Raymundo Marcial-Romero 2 and J. A. Hernández 2 1 Faculty of Computer Sciences, BUAP 2 Facultad de Ingeniería,

More information

Intelligent Agents. Pınar Yolum Utrecht University

Intelligent Agents. Pınar Yolum Utrecht University Intelligent Agents Pınar Yolum p.yolum@uu.nl Utrecht University Logical Agents (Based mostly on the course slides from http://aima.cs.berkeley.edu/) Outline Knowledge-based agents Wumpus world Logic in

More information

Harvard CS 121 and CSCI E-121 Lecture 22: The P vs. NP Question and NP-completeness

Harvard CS 121 and CSCI E-121 Lecture 22: The P vs. NP Question and NP-completeness Harvard CS 121 and CSCI E-121 Lecture 22: The P vs. NP Question and NP-completeness Harry Lewis November 19, 2013 Reading: Sipser 7.4, 7.5. For culture : Computers and Intractability: A Guide to the Theory

More information

Introduction to Solving Combinatorial Problems with SAT

Introduction to Solving Combinatorial Problems with SAT Introduction to Solving Combinatorial Problems with SAT Javier Larrosa December 19, 2014 Overview of the session Review of Propositional Logic The Conjunctive Normal Form (CNF) Modeling and solving combinatorial

More information

NP Complete Problems. COMP 215 Lecture 20

NP Complete Problems. COMP 215 Lecture 20 NP Complete Problems COMP 215 Lecture 20 Complexity Theory Complexity theory is a research area unto itself. The central project is classifying problems as either tractable or intractable. Tractable Worst

More information

NP-COMPLETE PROBLEMS. 1. Characterizing NP. Proof

NP-COMPLETE PROBLEMS. 1. Characterizing NP. Proof T-79.5103 / Autumn 2006 NP-complete problems 1 NP-COMPLETE PROBLEMS Characterizing NP Variants of satisfiability Graph-theoretic problems Coloring problems Sets and numbers Pseudopolynomial algorithms

More information

A Clause Tableau Calculus for MinSAT

A Clause Tableau Calculus for MinSAT A Clause Tableau Calculus for MinSAT Chu Min LI, a Felip MANYÀ b and Joan Ramon SOLER b a Université de Picardie, Amiens, France b IIIA-CSIC, Bellaterra, Spain Abstract. We define a clause tableau calculus

More information

Propositional Logic: Methods of Proof (Part II)

Propositional Logic: Methods of Proof (Part II) Propositional Logic: Methods of Proof (Part II) You will be expected to know Basic definitions Inference, derive, sound, complete Conjunctive Normal Form (CNF) Convert a Boolean formula to CNF Do a short

More information

An Efficient Decision Procedure for Functional Decomposable Theories Based on Dual Constraints

An Efficient Decision Procedure for Functional Decomposable Theories Based on Dual Constraints An Efficient Decision Procedure for Functional Decomposable Theories Based on Dual Constraints Khalil Djelloul Laboratoire d Informatique Fondamentale d Orléans. Bat. 3IA, rue Léonard de Vinci. 45067 Orléans,

More information

Propositional Fragments for Knowledge Compilation and Quantified Boolean Formulae

Propositional Fragments for Knowledge Compilation and Quantified Boolean Formulae 1/15 Propositional Fragments for Knowledge Compilation and Quantified Boolean Formulae Sylvie Coste-Marquis Daniel Le Berre Florian Letombe Pierre Marquis CRIL, CNRS FRE 2499 Lens, Université d Artois,

More information

Foundations of Artificial Intelligence

Foundations of Artificial Intelligence Foundations of Artificial Intelligence 8. Satisfiability and Model Construction Davis-Putnam-Logemann-Loveland Procedure, Phase Transitions, GSAT Joschka Boedecker and Wolfram Burgard and Bernhard Nebel

More information

Generating Models of a Matched Formula With a Polynomial Delay (Extended Abstract)

Generating Models of a Matched Formula With a Polynomial Delay (Extended Abstract) Generating Models of a Matched Formula With a Polynomial Delay (Extended Abstract) Petr Savický Institute of Computer Science The Czech Academy of Sciences Czech Republic savicky@cs.cas.cz Abstract A matched

More information

PROPOSITIONAL LOGIC. VL Logik: WS 2018/19

PROPOSITIONAL LOGIC. VL Logik: WS 2018/19 PROPOSITIONAL LOGIC VL Logik: WS 2018/19 (Version 2018.2) Martina Seidl (martina.seidl@jku.at), Armin Biere (biere@jku.at) Institut für Formale Modelle und Verifikation BOX Game: Rules 1. The game board

More information

arxiv: v1 [cs.lo] 8 Jun 2018

arxiv: v1 [cs.lo] 8 Jun 2018 Maximizing the Number of Satisfied L-clauses Mohamed El Halaby halaby@sci.cu.edu.eg Areeg Abdalla areeg@sci.cu.edu.eg Department of Mathematics, Faculty of Science, Cairo University Giza, Egypt arxiv:1806.02930v1

More information

SAT-Solving: From Davis- Putnam to Zchaff and Beyond Day 3: Recent Developments. Lintao Zhang

SAT-Solving: From Davis- Putnam to Zchaff and Beyond Day 3: Recent Developments. Lintao Zhang SAT-Solving: From Davis- Putnam to Zchaff and Beyond Day 3: Recent Developments Requirements for SAT solvers in the Real World Fast & Robust Given a problem instance, we want to solve it quickly Reliable

More information

Logic and Inferences

Logic and Inferences Artificial Intelligence Logic and Inferences Readings: Chapter 7 of Russell & Norvig. Artificial Intelligence p.1/34 Components of Propositional Logic Logic constants: True (1), and False (0) Propositional

More information

COMP3702/7702 Artificial Intelligence Week 5: Search in Continuous Space with an Application in Motion Planning " Hanna Kurniawati"

COMP3702/7702 Artificial Intelligence Week 5: Search in Continuous Space with an Application in Motion Planning  Hanna Kurniawati COMP3702/7702 Artificial Intelligence Week 5: Search in Continuous Space with an Application in Motion Planning " Hanna Kurniawati" Last week" Main components of PRM" Collision check for a configuration"

More information

Propositional Logic: Models and Proofs

Propositional Logic: Models and Proofs Propositional Logic: Models and Proofs C. R. Ramakrishnan CSE 505 1 Syntax 2 Model Theory 3 Proof Theory and Resolution Compiled at 11:51 on 2016/11/02 Computing with Logic Propositional Logic CSE 505

More information