(IJACSA) International Journal of Advanced Computer Science and Applications,

Vol. XXX, No. XXX, 2019

Hypercube graph decomposition for Boolean

simplification : an optimization of business process
verification

Mohamed NAOUM, Outman EL HICHAMI,

Mohammed AL ACHHAB, and Badr eddine EL MOHAJIR
New Technology Trends Team, Science and Technology Center for Doctoral Studies,
Abdelmalek Essaâdi University,
Tetouan, Morocco

Abstract—This paper deals with the optimization of busi- Model Checking (MC)[5], [9] and Automated Theorem
ness processes(BP) verification by simplifying their equivalent Proving (ATP)[10], [11].
algebraic expressions. Actual approaches of business processes
verification use formal methods such as automated theorem The MC approach uses software called model checker to
proving and model checking to verify the accuracy of the business exhaustively check whether an abstraction equivalent structure
process design. Those processes are abstracted to mathematical of the BP satisfies some properties expressed in temporal
models in order to make the verification task possible. However, logics. Simple Promela INterpreter (SPIN) is a widely used
the structure of those mathematical models is usually a Boolean model checker that verifies if a model writen in a C-like
expression of the business process variables and gateways. Thus modeling language called Process Meta LAnguage(Promela),
leading to a combinatorial explosion when the number of literals meets properties expressed as Linear Temporal Logic (LTL)
is above a certain threshold. This work aims at optimizing the
formulas[12], [13], [14]. Although this method has the advan-
verification task by managing the problem size.
tage of indicating the counter example violating the checked
A novel algorithm of Boolean simplification is proposed. propriety, it suffers from the state explosion problem[12] since
It uses hypercube graph decomposition to find the minimal its complexity is too high and the number of states grows
equivalent formula of a business process model given in its dis- exponentially.
junctive normal form(DNF). Moreover, the optimization method
is totally automated and can be applied to any business process The ATP (or automated deduction) is a subfield of math-
having the same formula due to the independence of the Boolean ematical logic dealing with automatic (or semi-automatic)
simplification rules from the studied processes. proving of mathematical theorems. The computer programs
This new approach has been numerically validated by com- allowing this task are called theorem provers[15].
paring its performance against the state of the art method Quine- First-order theorem proving is one of the most mature
McCluskey(QM) through the optimization of several processes subfields of ATP thanks to its expressivity that allows the
with various types of branching.
specification of arbitrary problems[16]. However, some state-
Keywords—Business process verification; Minimal disjunctive ments are undecidable[17] in the theory used to describe the
normal form; Boolean reduction; Hypercube graph; Karnaugh model. thereby, current research [18], [17], [19] deal with the
map; Quine-McCluskey challenge of finding subclasses of first-order logic(FOL) that
are suitable and decidable in the mapping of such models.
I. I NTRODUCTION Higher order logics are more expressive and can map wider
Business processes are key assets of any organization or range of problems than FOL, but theorem proving for these
information system[1], [2]. They are the communication inter- logics is not as developed as in the FOL[20].
face and the medium of exchange between the organization Regardless the used approach to verify a BP, its logical
stakeholders[3]. structure is deducted as a propositional logic formula written
BP describe the core business and govern the operation of in Disjunctive Normal Form(DNF)[2], [7]. The DNF can be
a system. Business Process Model and Notation (BPMN) reduced to a minimal form in order for the manipulation and
is the wide used standard for modeling BP in view of its practical implementation to become more efficient. Thus, an
simplicity and usability[4], [5]. Nevertheless, BP may contain optimization of the PB verification is achieved.
structural flaws[5] due to poor design or human errors. Hence, Since the simplification of Boolean expressions is exten-
the verification task is a crucial step between the modeling and sively used in the analysis and design of algorithms and logical
the execution phases of any BP. The complexity of real-life BP circuits, several methods were developed to perform this task:
and the use of automated modeling tools often lead to complex
models called “spaghetti” process models[6], [7] where manual − The algebraic manipulation of the Boolean expres-
verification is difficult to perform[8]. Therefore, automated for- sions aims at finding an equivalent expression by
mal methods are used instead. Automatic verification includes: applying the laws of Boolean algebra. However, for
www.ijacsa.thesai.org 1|Page
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. XXX, No. XXX, 2019

such methods, there is no fixed algorithm to be used

to minimize a given expression. Thus, choosing which
Boolean theorems to apply is left to the expert’s
ability.
− The Karnaugh map which is a pictorial and straight-
forward method[21]. First, a grid of the truth table
of the function to minimize has to be drawn. The
minterms of this grid have to be arranged in Gray code
which makes each pair of adjacent cells different only
by the value of one variable.
The problem is then converted into finding rectangular
groups of adjacent cells containing ones, these groups
should have an area that is a power of two (i.e., 1, 2,
4, 8. . . ). Consequently, unwanted variables are elimi-
nated. This method is easy to understand, however it is
a manual process which is not practical when dealing
with more than six variables [22].
− The tabulation method (also known as Quine Mc- Fig. 1. Main flow objects and sections flows of BPMN 2.0.
Cluskey algorithm)[23] is a useful minimization al-
gorithm when dealing with more than 4 variables. It
has a tabular form that makes it easy to implement The main role of the three categories above is to increase
in computer programs. It consists of finding all prime readability of the model without effecting its execution. There-
implicants of the function to minimize, and then tries fore the whole BP flow can be described with the remaining
to find the necessary ones that cover the function. two categories: Flow Objects and Connecting Objects[25].
Although this method is more practical than the previ-
ous ones, it is impaired by the redundancy during the The BPMN 2.0 specifies three Flow Objects: 1. Events,
search of prime implicants. Moreover, the application 2. Activities and 3. Gateways (see Fig. 1). These elements
of Petrick’s method[24] in a second phase is required are connected using Connecting Objects especially Sequence
to define essential prime implicants and resolve the flows.
cyclic covering problem. The Event elements indicate the various incidents that can
occur during the process execution. Three main type of events
This article introduces a novel technique to optimize the can be distinguished according to their trigger time: 1.Start
verification of a BP by simplifying its equivalent logical Events, 2.End Events, and 3.Intermediate Events. They indicate
formula written in the Disjunctive Normal Form (DNF). the beginning or the end of a process or simply any event that
This new simplification algorithm searches for the largest may arise in-between.
hypercubes of lower dimensions (called elements) that are
enough to cover all vertices in a partial cube graph mapping The Activity elements are used to indicate any performed
of the BP. A minimal equivalent DNF is then expressed as task in a process. Depending on the level of abstraction, an
a disjunction of the necessary hypercube abstractions in this Activity may be compound or atomic.
elements coverage. The Sequence flows are the arcs connecting related events
The rest of this paper is structured as follows: Section and activities. They define the chronological order of the
II describes how the BP is modeled in BPMN. Section III elements within a process. If the activation of a sequence flow
presents the main Boolean algebra simplification rules as well depends on some condition, then a Boolean variable is defined
as the hypercube properties that are used in the developed above it. Thus the immediate successor element is activated
algorithm. Section IV explains in details the simplification only if this condition is considered to be true.
algorithm and goes throw the used speedup tweaks. Our The Gateway elements are used to indicate any diver-
findings are presented and discussed in Section V . Finally, gence or convergence in a Sequence Flow. Depending of
a conclusion is given. their behavior, the five types of Gateways are: Exclusive,
Inclusive, Parallel, Event-Based, and Complex. They determine
II. B USINESS P ROCESS M ODELING AND N OTATION the branching, forking, merging, and joining of paths.
The most used business process modeling standard is Busi- The graph composed of Flow objects and their Sequence
ness Process Model and Notation(BPMN). It is a specification Flows connections describes the eventual executions of a BP.
of the Object Management Group (OMG)[25]. The modeling Each path of the graph going from the start to the end events
is done by interconnecting standard graphical symbols grouped indicates a single execution scenario. As an example, Fig. 2
in five categories: shows a simplified payment/delivery BP.
The Swimlanes and Artifacts categories are used to group Once the modeling of the BP is done, the designer must
objects into lanes and to provide additional descriptions. The choose which verification method to apply. The structure of
Data elements category is used to describe the flow of the the BP model is then extracted as a mathematical expression
data through the process. that depends on the used gateways and the sequence flows
www.ijacsa.thesai.org 2|Page
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. XXX, No. XXX, 2019

Table I. B OOLEAN ALGEBRAIC IDENTITIES

Accept Cash or
Check
Check or Cash
Addition A∨0=A A∨1=1 A∨A=A A∨A=1

Prepare package
multiplication A∧0=0 A∧1=A A∧A=A A∧A=0
identify the
for customer
payment method

Credit Card
Table II. B OOLEAN ALGEBRA PROPERTIES
Process Credit
Card Decision gateway

Addition (∨) Multiplication (∧)

Regular task
A∨B =B∨A A∧B =B∧A
Start event Sequence flow End event
A ∨ (B ∨ C) = (A ∨ B) ∨ C A ∧ (B ∧ C) = (A ∧ B) ∧ C
A ∧ (B ∨ C) = (A ∧ B) ∨ (A ∧ C)
Fig. 2. An Example of a Simple payment/delivery BP.

branching. The next section will present the necessary elements

used to map the logical structure of a BP and the main rules
used to simplify its equivalent formula.

III. B INARY REPRESENTATION AND REDUCTION RULES

A. Definitions
1) Boolean variable: A Boolean variable is a variable that
takes only one of the logical values: either 1 (meaning T rue)
or 0 (meaning F alse). The complement of a variable A is
denoted A and has the opposite value of A. A literal is either
the logic variable A or its complement A.
2) Minterm: A Minterm is a product (conjunction) of all
the variable literals. For instance, for three Boolean variables
A, B, and C the expressions ABC, A.B.C, and A ∧ B ∧ C
denote the same minterm. It means that C has the value 0 and
both A and B have the value 1. By assigning a power of 2 Fig. 3. Hasse diagram of the hypercube graph Q4
to each variable of a minterm Vn−1 ...V2 V1 V0 composed of n
variables Vi , the shorthand notation is md where d denotes
the decimal value of the binary expression Vn−1 ...V2 V1 V0 )2 . 2) Boolean algebra properties: In Boolean algebra, there
For example, m6 is the short hand notation of ABC because are three basic properties: commutative, associative, and dis-
110)2 = 6. tributive. Table II gives a summary of those properties:
3) Disjunctive Normal Form (DNF): A logical formula is 3) Boolean simplification rules: By using the identities
considered to be in Disjunctive Normal Form (DNF) if and and properties of Boolean algebra, a Boolean statement can
only if it is a disjunction (sum) of one or more conjunctions be simplified by reducing the number of literals using the
(products) of one or more literals [26]. A DNF formula following rules:
is in full disjunctive normal form if each of its variables
appears exactly once in every conjunction (minterm). The only ABC + ABC = BC (3)
propositional operators in DNF are and (denoted with . or ∧),
A + AB = A (4)
or (denoted with + or ∨), and not (denoted with ¬A or A).
The not operator can only be used as part of a literal, which A + AB = A + B (5)
means that it can only precede a propositional variable. The
following formula of three variables A, B, and C is in DNF: (A + B)(A + C) = A + BC (6)

f =AB C + AB C + AB C + AB C (1) C. The hypercube graph representation

A Boolean statement of n variables can be written in DNF
It can be written in shorthand notation as follow: with at most 2n minterms of n literals. By creating a vertex
for each minterm mi and linking each two vertices when their
f = m3 + m4 + m6 + m7 (2)
binary representations differ in a single digit (the Hamming
distance of their minterms is one), a hypercube graph (noted
B. Boolean algebra n-cube or Qn ) is created[27]. Fig. 3 gives a flat representation
of the hypercube graph Q4 .
1) Boolean algebra identities: In Boolean algebra, there
are four basic identities for addition (logical or) and four for A hypercube graph of n vertices can be viewed as the
multiplication (logical and) that holds true for all possible disjoint union of two hypercubes Qn−1 if an edge is added
values of a Boolean statement variables. Table I gives a from each vertex/minterm in one copy of Qn−1 to the corre-
summary of those identities: sponding minterm/vertex of the other copy. As shown in Fig. 4,
www.ijacsa.thesai.org 3|Page
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. XXX, No. XXX, 2019

m0

m1 m2 m4 m8

m6
Fig. 4. Construction of hypercube Qn from two Qn−1 hypercubes m3 m5 m10 m12

m9

the joining edges form a perfect matching between the blue

and black vertices.
m7 m11 m13 m14
In fact, every hypercube Qn of n > 0 is composed of
elements, or n-cubes of a lower dimension, on the (n-1)-
dimensional surface on the parent hypercube. The smallest
elements are the vertices (points). There is 2n of them. m15

In general, the number of m-cubes on the
boundary of a

n n n!


given n-cube is Em,n = 2n−m m where m = m!(n−m)! is Fig. 5. Reduction of a full DNF of 4 variables to hypercubes Q2 and Q3
the binomial coefficient.
A partial cube is an isometric subgraph of a hypercube. which gives an abstraction of the variable Vn . As a result,
The distance between any two vertices in the subgraph is the the hypercube Qn+1 gives an abstraction of n + 1 variables:
same as the distance between those vertices in the hypercube. n variables with the hypercube Q1n plus that of Vn .

Lemma III.1 Let Qn be a hypercube graph with n > 0 In the next section, an explanation of how the lemma III.1
minterms mi where i ∈ [0, 2n [. Let f be a DNF formula can be used as a key stone to perform the simplication of any
given by the disjunction of all Qn minterms. Then n variables formula written in DNF is given.
of f can be simplified. The abstracted equivalent formula is
easily obtained by identifying the common literals between IV. T HE S IMPLIFICATION ALGORITHM
the minterm with maximum shorthand notation value (denoted
In order to simplify a Boolean expression written in DNF,
mmax ) and the one with the minimum shorthand notation
its expression is represented as a partial cube P Qn of the
value (denoted mmin ). This abstraction is chosen to be called:
hypercube graph Qn , with n the number of variables in the
abstraction mmax with filter mmin .
DNF formula. The developed algorithm consists in finding the
largest elements (hypercubes) Qm , with m ≤ n, so that their
Proof: For instance, if n = 1 then Q1 is composed of disjoint union covers all vertices of the partial cube P Qn .
two minterms m0 and m1 of one variable v0 . By applying the The fewer is the number of necessary hypercubes Qm , the
identity v0 + v0 = 1, an abstraction of the variable v0 is given more abstract is the equivalent formula. As an example, the
(abstraction m1 with the filter m0 ). following DNF formula can be considered:
If n = 2 then Q2 is composed of four minterms f (A, B, C, D) = A B CD + A BCD + ABC D+
{m0 , m1 , m2 , m3 } each one is composed of two variables v0
ABCD + ABCD + AB CD + ABCD+
and v1 . By applying the same identity to two opposite sides
of Q2 an abstraction of the variables v0 and v1 is given (the ABC D + ABCD + ABCD (7)
abstraction m3 with the filter m0 ). In fact:
This formula is represented as a partial cube P Q4 with
f = m0 + m1 + m2 + m3 = v1 .v0 ∨ v1 .v0 ∨ v1 .v0 ∨ v1 .v0 vertices m1 , m3 , m4 , m5 , m7 , m9 , m11 , m12 , m13 , and
f = v1 .(v0 ∨ v0 ) ∨ v1 .(v0 ∨ v0 ) = v1 ∨ v1 = 1 m15 . Fig. 5 shows that the vertices of P Q4 (green and
yellow vertices) can be covered with the disjoint union of two
Let us assume that the lemma III.1 is correct for any hypercubes Q3 and Q2 .
n > 0. Let Q1n and Q2n be two hypercubes that their Using lemma III.1, three variables A, B, and C can
disjoint union form the hypercube Qn+1 . Each minterm mx = be reduced with the hypercube Q3 composed of vertices
mVn Vn−1 ...V2 V1 V0 )2 in Q1n forms a perfect matching with {m1 , m3 , m5 , m7 , m9 , m11 , m13 , m15 }. Thus Q3 is reduced to
another minterm my = mVn Vn−1 ...V2 V1 V0 )2 in Q2n . mx and m15 with the filter m1 which is equivalent to the expression D
my can be abstracted to mx because they differ by the value since it is the only variable that remains with the same value
of a single variable vn . In fact: in all minterms of Q3 (we have mmax = m15 = m1111)2 and
mmin = m1 = m0001)2 the abstraction is − − −1)2 ).

f = mx + my = Vn Vn−1 ...V2 V1 V0 ∨ Vn Vn−1 ...V2 V1 V0 The hypercube Q2 , composed of {m4 , m5 , m12 , m13 },
gives an abstraction of tow variables A and D. Thus Q2 is
f = (Vn ∨ Vn )Vn−1 ...V2 V1 V0 = Vn−1 ...V2 V1 V0 = mx reduced to m13 with the filter m4 which is equivalent to
www.ijacsa.thesai.org 4|Page
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. XXX, No. XXX, 2019

identify L the set of active RAM.

minterms
For each dimension n, with n ≤ 4, all formulas
1 2
were
3
tested
4
since there is only 65812 possible ones (22 +22 +22 +22 =
65812). For n > 4, the formulas to minimize were chosen up
no further reduction is is L
to x = 224 .
[Yes] No
possible empty ?
For each test, the running times, for both methods, were
pop into X the minterm
with maximum number of recorded starting from the feeding of the formula to minimize
set bits until the reception of the minimal equivalent DNF. The integer
no further reduction
from X is possible representing the input formula is then incremented for the next
identify S the set of x test. Since the execution time can vary significantly depending
[Yes]
active successors on the input size, we choose to plot the relative percent
is S
difference of the two algorithms runtimes. Each scatter in
empty ? Fig. 7 represents the result of one test that is given by the
delete from Y all R formula :
elements [No]
QM 0 s runtime − Our algorithm0 s runtime
100 ∗
pop into Y the farthest minimum of both runtimes
successor of X
update the best reduction no possible reduction A blue scatter indicates a result in favor of the proposed
for each minterm in R from X to Y algorithm while a red scatter indicates a result in favor of the
identify R the set of QM algorithm.
required intermediate
vertices The plot was generated using the python data visualization
possible mapping
from X to Y library Seaborn based on matplotlib.
are all R
[Yes] minterms [No] V. R ESULTS AND DISCUSSION
active?
From Fig. 7 we can conclude that our algorithm has better
performances than the QM Method since it has better results
Fig. 6. Organogram of the proposed Boolean reduction algorithm in 89.40% cases of the 224 conducted tests. Moreover, the
proposed algorithm is over 400% faster in 1380450 cases while
the QM method is over 400% faster only in 746 cases. Also
the expression BC (we have mmax = m13 = m1101)2 and this percent difference can reach over 2000% in 3829 cases in
mmin = m4 = m0100)2 the abstraction is −10−)2 ). favor of the developed algorithm and in no case in favor of
Finally, the disjunction of this two abstractions gives the the QM method.
minimal formula : One advantage of the new algorithm introduced in this
(A, B, C, D) = D + BC. (8) work, is that it follows a top-down approach: it searches first
for the largest hypercube that covers a minterm which means
If a vertex of the partial cube is covered by multiple hyper- that the algorithm does not waste time on smaller hypercubes
cubes, the largest one has to be considered. That way, each with less abstraction. In the counterpart, the Quine-McCluskey
vertex is surely covered with the most abstract expression. algorithm follows a down-top approach: it tries to find all
A simplified version of the reduction algorithm is sum- prime implicants of size 2 then size 4 and so on, which means
marized in Fig. 6. The algorithm starts with identifying the that it wastes time on multiple partial prime implicants before
vertices of the partial cube P Q that maps all the minterms of reaching the optimum formula.
the formula to minimize. Then, it tries to find, for each vertex A second advantage of the developed algorithm is that,
mi of the P Q, the largest hypercube (or hypercubes if there unlike for the tabular method, there is no need to use the
are many with the same size) that contains mi . Finally, the Pitrick’s method to solve the problem of cyclic covering. It
algorithm gives priority to external vertices then holds only the is simply solved by holding first the coverage of the external
necessary hypercubes to cover them all. The abstraction given vertices of the decomposition hypercubes.
by those hypercubes is the minimal equivalent expression of
the DNF to minimize. Finally, another advantage is the use of binary operations
that are directly supported by the microprocessor; it applies a
In the next section, the performance of the proposed simple binary and/or filters to find the successors of a given
algorithm will be compared with the Quine-McCluskey method vertex or to store the previous found coverage. For instance,
(QM). if there are six variables then there are 26 = 64 minterms,
All abstractions were performed using a Python implemen- instead of using a loop of 64 iterations, a single microprocessor
tation of the developed algorithm. They were then compared operation can be used to filter the active minterms in the partial
to the an optimized Python implementation of the standard cube.
Quine-McCluskey algorithm (This implementation is included
in the digital electronics simulation library BinPy). C ONCLUSION
The experiments were carried out on a conventional laptop Business Processes are indubitable tools for the modern
computer equipped with an Intel i5 processor and 8GB of business planning, but those models can include structural
www.ijacsa.thesai.org 5|Page
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. XXX, No. XXX, 2019

Fig. 7. Relative percent difference of the two algorithms’ runtimes.

flaws that are hard to detect with manual verification, which R EFERENCES
gives extreme importance to automatic verification. Formal [1] R. Heinrich, P. Merkle, J. Henss, and B. Paech, “Integrating business
methods verification algorithms suffer from the high complex- process simulation and information system simulation for performance
ity since the problem they try to solve is NP-hard, hence the prediction,” Softw Syst Mod, vol. 16, no. 1, pp. 257–277, 2017.
necessity to reduce the problem size by minimizing the number [2] D. Batory, “Feature models, grammars, and propositional formulas,” in
of literals. International Conference on Software Product Lines. Springer, 2005,
pp. 7–20.
In this paper, a novel technique of business processes [3] J. Stark, “Product lifecycle management,” in Product Lifecycle Man-
simplification has been presented. A simplication tool that agement. Springer, 2015, vol. 1, pp. 1–29.
performs literals reduction using hypercube decomposition has [4] H. Völzer, “An overview of bpmn 2.0 and its potential use,” in Inter-
been built. Moreover, the simplification algorithm was entirely national Workshop on Business Process Modeling Notation. Springer,
automated which makes the optimization task accessible to 2010, pp. 14–15.
the regular BP designers. Promising subject of research can [5] W. M. P. Van Der Aalst, M. L. Rosa, and F. M. Santoro, “Business
process management - don’t forget to improve the process!” Bus Inform
be explored in further depth, such as how machine learning Syst Eng, vol. 58, no. 1, pp. 1–6, 2016.
algorithms could be used to accelerate the simplification al-
[6] V. Gruhn and R. Laue, “Complexity metrics for business process mod-
gorithm, how the algorithm can be modified to reduce the els,” in 9th international conference on business information systems
spatial complexity, and finally, the possibility of adapting the (BIS 2006), vol. 85. Citeseer, 2006, pp. 1–12.
algorithm, view its characteristics, for quantum computing. [7] K. Batoulis, A. Meyer, E. Bazhenova, G. Decker, and M. Weske,
“Extracting decision logic from process models,” in International Con-
ference on Advanced Information Systems Engineering. Springer, 2015,
pp. 349–366.
[8] A. Förster, G. Engels, T. Schattkowsky, and R. V. D. Straeten, “Verifi-
cation of business process quality constraints based on visual process
patterns,” in First Joint IEEE/IFIP Symposium on Theoretical Aspects
of Software Engineering, TASE 2007, June 5-8, 2007, Shanghai, China.
IEEE Computer Society, 2007, pp. 197–208.
[9] A. Elgammal, O. Turetken, W.-J. van den Heuvel, and M. Papazoglou,
“Formalizing and appling compliance patterns for business process
compliance,” Softw Syst Model, vol. 15, no. 1, pp. 119–146, 2016.
[10] X. Tan, Y. Gu, and J. X. Huang, “An ontological account of flow-control
components in bpmn process models,” Big Data Inf Anal, vol. 2, no. 2,
pp. 177–189, 2017.
[11] S. Mallek, N. Daclin, V. Chapurlat, and B. Vallespir, “Enabling model
checking for collaborative process analysis: from bpmn to ‘network of
timed automata’,” Entrep Inf Syst - UK, vol. 9, no. 3, pp. 279–299,
2015.

www.ijacsa.thesai.org 6|Page
 (IJACSA) International Journal of Advanced Computer Science and Applications,
Vol. XXX, No. XXX, 2019

[12] E. Clarke, O. Grumberg, S. Jha, Y. Lu, and H. Veith, “Progress on the

state explosion problem in model checking,” in Informatics. Springer,
2001, pp. 176–194.
[13] Y. Li, A. Deutsch, and V. Vianu, “A spin-based verifier for artifact
systems,” Comput Res Rep, vol. abs/1705.09427, 2017.
[14] C. Wolter, P. Miseldine, and C. Meinel, “Verification of business
process entailment constraints using spin,” in international symposium
on engineering secure software and systems. Springer, 2009, pp. 1–15.
[15] L. C. Paulson, Isabelle: A generic theorem prover, ser. Lecture Notes
in Computer Science. Springer Science & Business Media, 1994, vol.
828.
[16] G. Buday, “Logic in computer science: Modelling and reasoning about
systems by huth michael and ryan mark, isbn 0 521 54310 x.” J Funct
Program, vol. 18, no. 3, pp. 421–422, 2008.
[17] S. Halfon, P. Schnoebelen, and G. Zetzsche, “Decidability, complexity,
and expressiveness of first-order logic over the subword ordering,”
in Logic in Computer Science (LICS), 2017 32nd Annual ACM/IEEE
Symposium on. IEEE, 2017, pp. 1–12.
[18] M. Elberfeld, M. Grohe, and T. Tantau, “Where first-order and monadic
second-order logic coincide,” ACM Trans Comput Logic, vol. 17, no. 4,
p. 25, 2016.
[19] M. Lamotte-Schubert, “Automatic authorization analysis,” Ph.D. disser-
tation, Saarland University, Saarbrücken, Germany, 2015.
[20] A. Gawanmeh and A. Alomari, “Challenges in formal methods for
testing and verification of cloud computing systems,” Scalable Comput
Pract Exp, vol. 16, no. 3, pp. 321–332, 2015.
[21] M. Karnaugh, “The map method for synthesis of combinational logic
circuits,” T Am Inst Elec Eng 1, vol. 72, no. 5, pp. 593–599, 11 1953.
[22] T. K. Jain, D. S. Kushwaha, and A. K. Misra, “Optimization of
the quine-mccluskey method for the minimization of the boolean
expressions,” in Fourth International Conference on Autonomic and Au-
tonomous Systems, ICAS 2008, 16-21 March 2008, Gosier, Guadeloupe.
IEEE Computer Society, 2008, pp. 165–168.
[23] W. V. Quine, “The problem of simplifying truth functions,” Am Math
Mon, vol. 59, no. 8, pp. 521–531, 1952.
[24] S. R. Petrick, “A direct determination of the irredundant forms of a
boolean function from the set of prime implicants,” AFCRC-TR-56,
vol. 10, p. 110, 1956.
[25] J. Mendling and M. Weidlich, Eds., Business Process Model and
Notation - 4th International Workshop, BPMN 2012, Vienna, Austria,
September 12-13, 2012. Proceedings, ser. Lecture Notes in Business
Information Processing, vol. 125. Springer, 2012.
[26] J. Cohen, “Review of ”introduction to lattices and order by b. a. davey
and h. a. priestley”, cambridge university press,” ACM SIGACT News,
vol. 38, no. 1, pp. 17–23, 2007.
[27] Y. Saad and M. H. Schultz, “Topological properties of hypercubes,”
IEEE T Comput, vol. 37, no. 7, pp. 867–872, 1988.

www.ijacsa.thesai.org 7|Page