Semantic features of processing hybrid dynamic workflows of design A N Afanasyev1, N N Voit2 and S Yu Kirillov3 1 Doctor of Engineering, First vice-rector, vice-rector of distance and further education, Ulyanovsk State Technical University, 32, Severny Venets, Ulyanovsk, 432027, Russia 2 Ph. D., associate professor of «Computer engineering», Ulyanovsk State Technical University, 32, Severny Venets, Ulyanovsk, 432027, Russia 3 PhD.student of «Computer engineering», Ulyanovsk State Technical University, 32, Severny Venets, Ulyanovsk, 432027, Russia Abstract. The article examines the semantic peculiarities of grammatical processing models of visual languages like RC ASCON-Volga, BPMN and eEPC as denotative and significative semantics. It is offered author's structure of the denotation and significata. Methods of control, analysis of qualitative and quantitative characteristics of workflows, transformation and interpretation of workflows are proposed by authors. A temporary automatic grammar of RC ASCON-Volga, BPMN and UML AD visual languages is proposed by authors for semantic processing of hybrid dynamic workflows, as well as for analysis and control of denotative and significative semantic errors. 1. Introduction The paradigm associated with the hybrid dynamic nature of their nature is increasingly dominating in the creation of project workflows. Hybridity is defined not only as the development of models using different diagram bases (for example, UML AD [1], BPMN [2], IDEF0 [3]), but also as the composition of orchestration and choreography [4, 5] in the form of an ensemble. Dynamism is determined by the need for immediate response to emerging production requests and contains the concept of «time», so the elimination of errors in the flow of work is a significant scientific and technical problem. Under treatment refers to the analysis, control, transformation, and interpretation workflows, and analysis and control diagrammatically models of the error associated with denotative and significative the semantics. Denotative semantics of diagram models is represented by a sequence of temporal words of the formal automaton language in the form of traces and defines antonymy, synonymy of these words in order to identify errors in the events of diagram models. The significative semantics of diagram models reveals the relations of isomorphism, homomorphism of these traces for the purpose of localization of structural errors in diagram models, and also for their subsequent transformation. Changing the design of the diagram models is possible with the identified synonyms, antonyms of graphic words, isomorphic traces and the possibility of combining such traces into a single track. A necessary condition for the transformation of the flow of design work is the presence of a semantic error. Detection of such errors is possible with the help of step-by-step automated interpretation (tracing) of the project work flow in debug mode. Automata-based grammars allow you to present a diagram model of the flow of design work in the form of a graph with vertices, arcs and in a visual form to represent the process of interpreting the flow of design work as a system of transitions. The methods of analysis can be used to study the qualitative and quantitative characteristics of project work flows. Under qualitative characteristics refers to the logical-algebraic correctness of workflows, formalized using graph theory, networks, workflows, matrix matching, graphical modeling languages, including Unified Model Language, Business Process Management Notation, IDEF0 and eEPC, etc., and also the evolutionary approach, logic statements, etc. Quantitative characteristics represent the effectiveness of the execution of the workflow in the 355 parameters, such as average service time, the utilization rate of production capacity (downtime), etc. The efficiency of workflows is evaluated using simulation modeling (Petri nets), Markov chains and Queuing theory (Queuing systems). In this paper, the authors developed a temporary automatic grammar of visual languages of RC ASKON-Volga, BPMN and eEPC, as well as denotative and significative semantics of diagram models of visual languages as a general structure for semantic processing of diagram models of hybrid dynamic design workflows. Processing will reduce the time, improve the success and quality of charting models. The mathematical apparatus of processing of diagram models of hybrid dynamic flows of design work allows to simulate the process of workflows in the form of a finite state machine. The work contains an introduction, the author's description of denotative and significative semantics of workflows in the visual languages of RC ASCON-Volga, BPMN and eEPC. The review of qualitative and quantitative methods for estimating the characteristics of workflows presents mathematical tools for the analysis, control and modeling of workflows. In the section of the Temporal automaton RVTI- grammar author presents automata-based temporal grammar for visual language RC ASKON-Volga, BPMN and UML AD. In the section Transformation the author's method of transformation of diagrammatic models of workflows by means of the developed author's temporal automatic grammar is offered. In conclusion, briefly concludes the work and identifies future directions of work. 2. Related work The authors investigate some works that consider the specification of document flow, verification and translation. Several papers focused on the definition of formal semantics and validation methods for workflows using Petri nets, process algebra, abstract state machine, see for example [6-16]. In [12, 13], Decker and Weske propose a Petri net-based formalism for determining choreographies, properties as realizability and local applicability, and a method for verifying these two properties. However, they consider only synchronous communication and does not explore the association with languages modeling of interaction of a high-level BPMN. Bultan and Fu [17] determine a sufficient condition for analyz-ing the feasibility of choreographies defined using UML collaboration diagrams (CD). In [18], Salaün and Bultan modify and extend this work with the feasibility analysis method by adding a synchronization message among peers. This method controls the realizability of CDs for bounded asynchronous communication. The feasibility problem for Message sequence diagrams (MSCs) has also been studied (e.g. [19, 20]). In [20], the authors offer bounded MSCS graphs which are bound-ed by BPMN 2.0 because branching and looping behavior are not supported by CDs and MSCs (there is no selection in CDs, there are no some looping behaviors in MSCs, and only Self- loops in CDs). In [21] BPMN behavior is studied from the semantic point of view and several BPMN patterns are proposed. This work is not theoretically justified and is not complete, it discusses only some of the laws. Lohmann and Wolf [22] propose to analyze existing patterns and control them with compatible patterns. In [23], the authors focused on the translation of BPMN into the algebra of processes for the analysis of choreography using model checking and equivalence. The main limitation of these methods is that they do not work when there are different types of diagrams at the same time, which means that in some cases the input diagrams cannot be analyzed. There are the following paradigms of analysis and control of quality characteristics of work flows: model checking; equivalence check; deductive verification (Prolog language). The model checking approach is intended for the analysis, control of workflows by means of formal check of whether the given logical formula is executed on the given structure (whether the given logical formula f will be true for the given system of transitions M, i.e. whether M will be model f). The main disadvantage of the campaign is the study of the model, not the system itself, so the question arises about the adequacy of the model to the system, and the complexity of the solution of the above problems is exponential. Deductive verification involves checking the correctness of the workflow, which is reduced to proving theorems in a suitable logical system with the help of axioms and inference rules (for example, with the help of the Prolog language, automatic grammars, etc.). This very complex procedure can not be fully automated, it requires the participation of a person acting on the basis of assumptions and guesses, using intuition in the construction of invariants and non-trivial choice of alternatives. The equivalence test determines the equivalence of formal models of specification, implementation and execution 356 (behavior) of workflows based on the algebra of processes (Calculus of interacting systems). Simulation modeling is a flexible approach to analysis and is applied almost always to the analysis of workflows, which is reduced to determining the path in the reachability graph, taking into account the probability of distribution. Multiple execution of workflows using a computer provides «ease» of understanding of the functional people who do not have mathematical training. Visual representation and analysis of workflows is available in many tools for modeling workflows. Usually use Petri nets in the modeling and analyzed the following properties: reachability (reachability) – which sets out that the final state of the system is reached when any sequence of transitions from positions i. This property also implies that upon reaching the final position of the network there are no chips in the intermediate positions; security (safety), establishes that the processes do not exist hangs (deadlocks), looping dead ends; vitality (liveness) – specifies that the system does not contain unnecessary items that will never be fulfilled. The lack of liveliness means either redundancy of the business process in the designed system, or indicates the possibility of loops, deadlocks, locks. Flow rate, transmission rate, waiting time, service time, and capacity utilization can be calculated using queue theory. If we are interested in the formation of a separate queue to multiple resources of the same type, it is necessary to confine the system with a single queue. When considering the entire flow, Queuing systems are best used. The main models used in Queuing theory are single-and multi-channel Queuing systems (QSOS). The most simple model of the workflow to determine complexity of the corresponding phase can be obtained if we accept the assumption about the absence of consequences in the process, meaning that the next job in the flow depends only on the current state and does not depend on previous States. In this case, the flow of work becomes a Markov process determined by a variety of inherent conditions and the matrix of transition probabilities, and a probability distribution of states in the initial moment of time. 3. Denotative and significative semantics of diagram models of visual languages of RC ASCON- Volga, BPMN and eEPC Denotative semantics [24-26] of any visual language is represented by denotates in the form of graphical objects (words). The denotate of a word in the theory of visual languages is understood as an instance of a class with specific values of properties characterizing the belonging of a word to a subject. The authors have developed a general structure of the denotation and significata graphic words. The general structure of the graphical class instance of a word is displayed in listing 1. Listing 1. Generalized structure of the graphic denotation of the word. The class name class=start beginning property 1=value 1 property 2=value 2 property n=value n end The structural units of the parameters (properties) of the graphical word represent the signature of the word, and the properties themselves are combined into a class (listing 2). Listing 2. Generalized structure of significata graphic words. Class name: text beginning property 1: type 1 property 2: type 2 property n: type n end The description of denotates and significations in the set-theoretic language has the following form. , (1) 357 where is the name of the denotation (class instance) the graphic word; is the value of the first property of the graphic word; is the value of the second property of the graphic word; is the value of the third property of the graphic word; is the value of n-th property of a graphical word. , (2) where is the name of significata (class) graphic words; is structure of the first property of the graphic word; is structure of the second property of the graphic word; is structure of the third property of the graphic word; is structure of n-th property of a graphical word. The difference between the significate and the denotate is that the denotate is an instance of the significate, i.e. the properties (structure) of the denotate have specific values. Further in tables 1, 2, 3 denotative and significative semantics of grammatical models of hybrid dynamic streams of design works of visual languages of RC ASKON Volga, BPMN, eEPC are presented [27-35]. Table 1. Denotative and significative semantics of the visual language of RC ASCON-Volga. Graphic Concept Class Class The values of class properties word properties A collapsed subprocess that can be called multiple times. Procedure Input Has only one inbound link of type «go to procedure» Algorithm Input=work thread Output Algorithm=ishodnyh Output=Potocari Collapsed subprocess, the action is required to be Task Input Input=work thread performed by the user Text of task Textpane=scriptside Output Output=work thread Collapsed subprocess, the implementation of which is Iteration Algorithm Algorithm=source code required repeatedly Used in conjunction with «go to procedure» and Procedure call Adres call Adres call procedure=number «procedure» block procedure Used in conjunction with «go to procedure» and Thread creation Name Name=stream_name «procedure» block Adresie Adresie=number The operation performed by the user No transit opcode opcode=number operand address 1 operand address 1=number operand address 2 operand address 2=number Automated (automatic) execution of operations The script (auto Algorithm Algorithm=source code operation) It has only two outgoing connections. True and False, Branching Condition Condition=source code respectively Allows you to connect parts of the chart at different Phantom From level Isorena=numerowana levels of nesting. In fact a connection Against the level Karouny=numerowana Used in conjunction with «waiting». Three incoming, Event Time Time=number one of them «Synchro action», notifying about the event. Used in conjunction with «waiting». Three incoming, Semaphore Beginning Start=number one of them «Synchro action», notifying about the Completion Completion=number beginning and completion of events. Has two outgoing branches, one of them «Synchro Activate Output Output=Boolean action», notifying «Event» on successful completion of the event It has two outgoing branches, one of them «Synchro Expectation Duration Duration=number action», which monitors the status of the event. Skips the stream only if the event succeeds It has two outgoing branches, one of them «Synchro Increment With C=the number action», notifying «Semaphore» about the beginning of On By=number the event execution Step Step=number It has two outgoing branches, one of them «Synchro Decrement From From=number action», notifying «Semaphore» about the completion of To Up=number the event Step Step=number 358 Table 2. Significative and denotative semantics of the visual language of BPMN. Graphic Concept Class Class The values of class word properties properties Has no incoming threads Starting event Output Output=work thread Only one incoming and outgoing stream Intermediate event Input Input=work thread Operation Operation=algorithm Output Output=work thread Can be a trigger Intermediate event Input Input=work thread «Message» Event Event=algorithm Output Output=work thread Has no outgoing streams Final event Input Input=work thread Incoming stream Action Input Input=work thread Procedure Procedure=algorithm Output Output=work thread Only one outgoing branch is allowed to flow Exclusive gateway Input Input=work thread Condition Condition=algorithm Output Output=work thread It can be only one trigger event Gateway the Event- Input Input=work thread based Event Event=algorithm Output Output=work thread Activated only if there is a thread on each incoming Parallel gateway Input 1 Input 1=work stream branch Input 2 Input 2=work stream Input 3 Input 3=work stream Input s Input=work thread Output Output=work thread Can connect to any flow element by association Data object Document Document=text The relationship between the graphical objects A regular relationship From From=graphical object To To=graphical object Allows you to monitor an invalid sequence of elements Event-based gateway From From=graphical object communication To To=graphical object Associative relationship between graphical objects by key Association From From=graphical object To To=graphical object Table 3. Significative and denotative semantics of the visual language eEPC. Graphic Concept Class Class The values of word properties class properties It is the fact of accomplishment of something, and Event Time Number not having duration in time, or this time to aspire to zero (or does not matter) Execution of a function always ends with the event Function Function The function=algorithm Workflow direction Through a process From From=work thread To To=work thread The object reflects the various organizational units Unit Name Name=text of the enterprise A logical operator that defines the relationship exclusive or From=work thread between events and functions within a process. From To=work thread Allows you to describe the branching process To A logical operator that defines the relationship OR From=work thread between events and functions within a process. From To=work thread Allows you to describe the branching process To A logical operator that defines the relationship AND From=work thread between events and functions within a process. From To=work thread Allows you to describe the branching process To The object reflects the media Information (Material) Document Document=text Core business process Main process From To From=graphical object To=graphical object Structural and functional element Component Name Name=text Area of research Subject area Name Name=text A set of simple processes, collected on the same Process group Name Name=text basis and solve the same problem The relationship between the graphical objects Dynamic connection line From To From=graphical object To= graphical object 359 Semantic errors of grammatical models of work flows include the following errors [36]. Synonym mismatch (denotative error). Temporal words of visual language and are synonyms if and only if and indicated synonymy of the words as . The identical equality of the word determines the similarity (similarity) of the structure and values of denotate features. A mistake is the situation when the name of the denotates of words in two temporal traces of the graphical language are similar, but the values of other features are very different. In practice, this situation is presented as follows: the analysis of the grammatical model of the visual language reveals the structural similarity of words and names of denotates, but the values of other features of denotates of words are different. To present variants of the composition of products under different conditions: in versions, substitutability and interchangeability, in this situation, the implementation of interchangeability of such words in the diagram model of the visual language is a mistake of non-conformity of synonyms. Discrepancy of antonyms (denotative error). Temporal words of visual language and are antonyms if and only if and is denoted by antonyms to words as . Identical to the opposite of the two words defines the similarity (similarity) of the structure and the opposite (inversely) characteristic value of the denotation. Generally, the words «Beginning» and «End» in charting are antonyms of the graphical language. A mistake is the project situation when the name of denotates of words in two temporal traces of graphic language are opposite (inverse), but the values of other signs are very similar. In practice, this situation is presented as follows: when analyzing the diagram model of the visual language, structural similarity of words and inversion of denotate names are revealed, but the values of the other signs of denotate words are similar. In this situation, the interchangeability of such words in the diagram model of the visual language is a mistake of discrepancy of antonyms. Conversionist relations is to bind antonyms diagrammatically models of visual languages that describe the same design situation, but with different roles. Error conversionist relations is significative, i.e. structural (structural), and is defined as the lack of these relations between antonyms diagrammatically models describing the same design situation, but with different roles. The inconsistency of the objects is significatively mistake. Is the absence of a relationship between dependent temporal words. 4. Temporal automata-based RVTI grammar The temporal automaton RVTI grammar of the language L (G) is called the ordered eight of non- empty sets   , where V   e , e  1.L is auxiliary alphabet (the alphabet of operations on internal memory, represented by the store or elastic band); is alphabet of graphic symbols (objects); is quasi-terminal alphabet, which is an extension of the terminal alphabet ∑; is clock ID (counter); is a lot of timestamps, and ; is temporal aspect ratio , where variable c (ID hours), the ratio ), describing the condition of the occurrence of an event t l ; is scheme of grammar G (the set of names of products of complexes, each complex ri consists of a subset products' ); is the axiom of RVTI-grammar (name of the initial complex of products), is the final set of products. Product have the form , where is n-ary ratio, which determines the type of operation on the internal memory, depending on (1 – write, 2 – read, 3 – a comparison), and ; is words as a pair of quasi-symbol and timestamp; is name of the successor product complex. The language of this grammar contains words of the form and and and , represents the alignment . In tables 4 and 5 present the temporal grammars for specific language RC ASKON Volga and language BPMN. 360 Table 4. Temporal RVTI grammar language RC ASKON Volga. Complex-source Quasi term The complex receiver The memory operation r0,t0 A0 r3,t3  /E r1,t1 return r2,t4 w2(b4m) r2,t2 vA r1,t1 w1(s1m,t4m), CALL vA/E vIT r1,t1 w1(s1m,t4m), CALL vIT/E Ak r4,t4  Akm r5,t5 w1(1t(1),it(2))/w2(et(1))/E _Akm r5,t5 w1(inc(mt(1)))/w3(mt(1)