Knowledge Based UML Information Flow Model Transformation Algorithm Ilona Veitaite Audrius Lopata Institute of Applied Informatics Institute of Applied Informatics Vilnius University, Kaunas Vilnius University, Kaunas Faculty Faculty Kaunas, Lithuania Kaunas, Lithuania Ilona.Veitaite@knf.vu.lt Audrius.Lopata@knf.vu.lt Abstract – The main scope is to present UML Information composition. Enterprise model stores knowledge that is Flow model generation from Enterprise model (EM) necessary for IS development process only and will be used transformation algorithm. The transformation algorithm during all phases of IS development life cycle [7, 14, 15, 16]. description is presented in details and depicted steps. Whole generation process steps are illustrated by particular example There is given formalized Enterprise meta-model following the transformation algorithm step by step. description, which is needed to define UML Information Flow model generation process algorithm. Enterprise model can be Keywords – Enterprise Model, Knowledge-based, IS Engineering, described as Malcev algebra based algebra system (Fig. 1) [10, UML, Information Flow, Information Item. 19]: M1= (1) I. INTRODUCTION where M1 – Enterprise model as algebra system; K – There have been quite many attempts for the analysis of elements set of M1 system; K={K1, K2,…, K21}, where UML models generation from different knowledge based K1,....K21 EM meta-classes; R – set of relationships between models combining other modelling languages, workflow elements, where R={r1, r2, r3}. patterns and frameworks or even generation from natural language specifications [1, 2, 8]. For each set of K element Kn composition is defined as: Kn=<{an1, an2,…,ank}, {mn1, mn2,…,mnl}>, where {an1, UML models are receiving an increasing attention from an2,…,ank} – attributes of Kn element, {mn1, mn2,…,mnl}– researchers in the recent years. It is a very challenging target methods of Kn element. for analysis of UML models since the knowledge about an enterprise system is allocated within several model views. Enterprise model M1 composition is as follows: UML models are maintained to decrease the confusion of the problem with the increase enterprise changes. By operating M1=<{K1, K2,...,K21}, {r1, r2, r3}> (2) UML models knowledge can be effectively expressed and can be used simply in all phases of IS development life cycle [2, 6, 9, 18]. UML as one of the main components of IS development life cycle phase models, can be generated in semi-automatic way from knowledge repository – Enterprise model. This kind of realization will improve the efficiency of these participants of information system development process: system analyst and/or system designer and/or system developer. II. ENTERPRISE MODEL ELEMENTS ROLE VARIATIONS Enterprise meta-model is formally determined enterprise model composition, which contained of a formalized enterprise model alongside with the general principles of control theory. Fig. 1. An Enterprise meta-model graphical schema based on Malcev Enterprise model is the main source of the requisite knowledge algebra [10,19] of the specific problem domain for IS engineering and IS reengineering processes [3, 4, 5, 14, 23]. where: K1 – meta-class Process, K2 – meta-class Function, Enterprise meta-model manages Enterprise model K3 – meta-class Actor, K4 – meta-class Event, K5 – meta-class Goal, K6 – meta-class Material Flows, K7 – meta-class Input Material Flow, K8 – meta-class Output Material Flow, K9 – Copyright held by the author(s). meta-class Information Flow, K10 – meta-class Interpretation, K11 – meta-class Data Processing and Solution Making, K12 – 30 meta-class Realization, K13 – meta-class Information Activity, K14 – meta-class Business Rules, K15 – meta-class Interpretation Business Rules, K16 – meta-class Data Processing and Solution Making Business Rules, K17 – meta- class Realization Business Rules, K18 – meta-class Process Output, K19 – meta-class Information processing Input Attributes, K20 – meta-class Information processing Output Attributes, K21 – meta-class Process Input, r1 – Aggregation, r2 – Generalization, r3 – Association. Information systems design methods indicates the disposition of systems engineering actions, i.e. how, in what order and what UML model to use in the IS development process and how to implement the process (Table 1). Majority of them are based on different types of models describing varying aspects of the system qualities. Meaning of each model can be defined individually, but more important is the fact that each model is the projection of the system. An inexperienced specialist can use UML models inappropriately and the description of the system will possibly be insufficient or even mistaken [11, 12, 13, 20]. Fig. 2. The top level transformation algorithm of UML models generation TABLE I. ENTERPRISE MODEL BUSINESS RULES ELEMENTS ROLE from EM process VARIATIONS IN PART OF UML DYNAMIC MODELS EM UML Model element UML Dynamic Model Transformation algorithm of UML models generation from Enterprise model is top level algorithm for enterprise meta- Extend Use Case Model model based UML model generating process (Fig. 2). Main Include Use Case Model steps for generating process are identifying and selecting UML model for generating process, identifying starting (initial) Association Use Case Model element for the selected UML model and selecting all elements related to this UML model, generating enterprise model Busines Rule Control Nodes Activity Model elements to UML model elements and generating the whole Time Constraint Timing Model UML model. Destruction Occurrence Timing Model Table 2 presents UML Information Flow model elements … … and its descriptions. Pseudostate State Machine Model TABLE II. UML INFORMATION FLOW MODEL ELEMENTS DESCRIPTIONS … … [17, 21] Determining specific UML model and selecting the initial UML Description Information model element is reasonably meaningful, because further Flow Model generating process relies on it. Many UML model elements element iterates in different UML model, but these elements describe A dynamic classifier which specifies a role played different aspects of the system. In example Enterprise model by an external entity that interacts with the subject element Business rule has different signification in different Actor (e.g., by exchanging signals and data), a user of the UML models [11, 12, 13, 20]. designed system, some other system or hardware using services of the subject. A classifier which describes a set of objects that III. TRANSFORMATION ALGORITHMS Class share the same: features, constraints, semantics (meaning). When Enterprise model as enterprise knowledge storage A directed relationship that is used as a allows – the stored knowledge is sufficient, validated and Information specification of some kind of “information verified – to generate UML models applying transformation flow channel” for unidirectional transmission of algorithms. This kind storage can be used not only for information from sources to targets. A classifier which represents some information knowledge of the enterprise gathering, but also as a tool that transferred within a system from source(s) to minimizes IS reengineering volume of work if any changes Information target(s) of information flow and provides no occur in an enterprise. UML models generation from Enterprise item details about the information they transfer as they model is implementation of knowledge based IS development do not have features. life cycle design phase. 31 There is given formalized UML Information Flow model entities at some high levels of abstraction and it is directly description. UML Information Flow also can be described as related with UML Class and Use Case models (Fig. 4). This Malcev algebra based algebra system (Fig. 3) [10,19]: model describes information flows and provides information to Class and Use Case models. M4= (3) where M4 – UML Information Flow model as algebra system; K – elements set of M4 system; K={K34, K35,…, K37}, where K34,....K37 UML Information Flow meta-classes; R – set of relationships between elements, where R={r1, r2, r3}. UML Information Flow M4 composition is as follows: M4={K34, K35,…, K37},{r2},{r3}> (4) where: K34 – meta-class Actor, K35 – meta-class Class, K36 – meta-class Information Flow, K37 – meta-class Fig. 4. UML 2.5 Models Overview fragment [17, 21] Information Item, r2 – Generalization, r3 – Association. Information flows can be useful to describe circulation of information through a system. These flows represents aspects of models not yet completely specified or with less details. Fig. 3. UML Information Flow graphical schema based on Malcev algebra According to the figure 3 it is clear that Enterprise model elements: Actor, Process, Function, Information Flow, Information processing Input Attributes, Information processing Output Attributes can be generated as UML Information Flow model elements: Actor, Class, Information Flow, Information Item. TABLE III. INTERSECTION BETWEEN ENTERPRISE MODEL AND UML INFORMATION FLOW MODEL ELEMENTS Enterprise model set UML Information Flow Formal element model set element description Actor (K3) Actor (K34) φ1: K3→K34 Process (K1) Class (K35) φ1: K1→K35 Information processing Information Item (K37) φ2: K19→K37 Input Attributes (K19) Information processing Information Item (K37) φ2: K20→K37 Output Attributes (K20) Information Flow (K9) Information Flow (K36) φ6: K9→K36 Fig. 5. Transformation algorithm of UML Information Flow model Table 3 presents intersection between Enterprise model and generation from EM process UML Information Flow model elements, where formal description of Enterprise model elements generated to UML Transformation algorithm of UML information Flow model Information Flow model elements according to Malcev algebra generation from Enterprise model process is presented in the can be found. figure (Fig. 5) and is illustrated by following steps:  Step 1: According to the top level transformation A. UML Information Flow Model Transformation Algorithm algorithm of UML models generation from EM UML Information Flow Model belongs to dynamic UML process, UML Information Flow model is identified models part and shows exchange of information among system 32 for the generation process. So the initial element for clinic to veterinary appointment in order to get the ultrasound UML Information Flow model is Actor element. examination, surgeon evaluation and veterinary consultation.  Step 2: UML Information Flow model Actor element Firstly, pet owner registers his pet in Veterinary clinic is generated from Enterprise model. registration system, orders the ultrasound examination in  Step 3: Process element from Enterprise model which ultrasound information system, then follows the process of the is related with the initial actor element is selected. examination, data storage and examination data sending to surgeon, surgeon analyses examination data and writes the  Step 4: UML Information Flow model Class element diagnosis using reviewing and evaluating system and sends it to is generated from Enterprise model. veterinary through the reviewing and evaluating system, who  Step 5: Information Flow element as link of other gives the result to pet owner. elements from Enterprise model which is related with the process element is selected. Detailed stages of Veterinary clinic example processes  Step 6: UML Information Flow model Information stored in Enterprise model are described: Flow element as link of other elements is generated  Stage 1 – Pet owner registers his pet in veterinary from Enterprise model. clinic registration system. Information system  Step 7: Information processing Input Attributes manages pet owner registration and services ordering element as definition of link element from Enterprise and is responsible for updating information. model which is related with the process element is  Stage 2 – Pet registration information from veterinary selected. clinics registration system is connected to ultrasound  Step 8: If UML Information Flow model Information examination registration system. System manages item element is definition of link to next element then examination order scheduling. it is generated from Enterprise model.  Stage 3 – Data gaining system acquires and creates  Step 9: Else Information processing Output Attributes medical data while a pet is present (in example: element as definition of link to previous element from ultrasound, tomography etc.) Enterprise model is selected.  Stage 4 – Data storage system manages examination  Step 10: UML Information Flow model Information data storage and sharing inside Veterinary clinic. item element as definition of link to previous element  Stage 5 – Surgeon gets data from data storage system, is generated from Enterprise model. evaluates it through reviewing and evaluating system  Step 11: UML Information flow elements Information and prepares diagnosis response. item and Information Flow are linked.  Stage 6 – Veterinary gets diagnosis response prepared  Step 12: UML Information flow elements Information by surgeon through reviewing and evaluating system. Flow and Class are linked.  Stage 7 – Pet owner gets diagnosis information during  Step 13: There is checking if there are more the appointment with veterinary. Information flows in Enterprise model related to UML Transformation algorithm of UML Information Flow model Information Flow model. In case, there are, algorithm generation of stage 1 of Scheduled workflow for Ultrasound goes back to step 5. examination for the pet in Veterinary clinic example from  Step 14: UML Information flow elements Class and Enterprise model process is illustrated by following steps: Actor are linked.  Step 1: Selected initial element for UML Information  Step 15: There is checking if there are more Processes Flow model is Actor element. in Enterprise model related to UML Information Flow  Step 2: UML Information Flow model Actor element model. In case, there are, algorithm goes back to step is generated from Enterprise model, in certain 3. example first actor is Pet owner.  Step 16: UML Information flow element Actor is updated. First two steps of transformation algorithm is presented in  Step 17: There is checking if there are more Actors in table 4. Enterprise model related to UML Information Flow model. In case, there are, algorithm goes back to step TABLE IV. STEP 1 AND STEP 2 IN UML INFORMATION FLOW MODEL 1. GENERATION PROCESS  Step 18: Else all UML Information Flow model Generated elements and links are generated from Enterprise Enterprise UML Model. Transformation algorithm part model Information element Flow model B. Generated UML Information Flow Model Example element Generation of UML Information Model is illustrated with the example of Scheduled workflow for Ultrasound examination for the pet in Veterinary clinic [20, 21]. Information of this example is stored in Enterprise model. Example shows, how pet owner registers his pet in veterinary 33  Step 3: Process element from Enterprise model which TABLE VII. STEP 7, STEP 8 AND STEP 9 IN UML INFORMATION FLOW MODEL GENERATION PROCESS is related with the initial actor element is selected.  Step 4: UML Information Flow model Class element Generated UML is generated from Enterprise model, in certain Transformation Enterprise model Information example first class is Pet registration. algorithm part element Flow model element Other two steps of transformation algorithm is presented in table 5. TABLE V. STEP 3 AND STEP 4 IN UML INFORMATION FLOW MODEL GENERATION PROCESS Generated UML Transformation Enterprise model Information  Step 10: UML Information Flow model Information algorithm part element Flow model item element as definition of link to previous element element is generated from Enterprise model.  Step 11: UML Information flow elements Information item and Information Flow are linked. Next two steps of transformation algorithm is presented in table 8.  Step 5: Information Flow element as link of other elements from Enterprise model which is related with TABLE VIII. STEP 10 AND STEP 11 IN UML INFORMATION FLOW MODEL GENERATION PROCESS the process element is selected.  Step 6: UML Information Flow model Information Generated UML Transformation Enterprise model Information Flow element as link of other elements is generated algorithm part element Flow model from Enterprise model, in certain example first element Information flow is between Pet owner and Pet registration. Other two steps of transformation algorithm is presented in table 6. TABLE VI. STEP 5 AND STEP 6 IN UML INFORMATION FLOW MODEL GENERATION PROCESS Transformation Enterprise Generated UML  Step 12: UML Information flow elements Information Information Flow Flow and Class are linked. algorithm part model element model element Step 12 of transformation algorithm is presented in table 9. TABLE IX. STEP 9 IN UML INFORMATION FLOW MODEL GENERATION PROCESS Generated UML Transformation Enterprise model  Step 7: Information processing Input Attributes algorithm part element Information Flow model element element as definition of link element from Enterprise model which is related with the process element is selected.  Step 8: If UML Information Flow model Information item element is definition of link to next element then it is generated from Enterprise model in certain  Step 13: There is checking if there are more example first Information item is Pet information. Information flows in Enterprise model related to UML  Step 9: Else Information processing Output Attributes Information Flow model. In case, there are, algorithm element as definition of link to previous element from goes back to step 5. All steps form the 5 are repeated. Enterprise model is selected. Next two (in other case three) steps of transformation Step 13 of transformation algorithm is presented in table 10, algorithm is presented in table 7. showing the result after repetition steps from step 5. 34 TABLE X. STEP 13 IN UML INFORMATION FLOW MODEL GENERATION PROCESS Transformati Enterprise Generated UML Information on algorithm model Flow model element part element  Step 14: UML Information flow elements Class and Actor are linked, in certain example Pet owner is linked to pet registration. Step 14 of transformation algorithm is presented in table 11, showing the result after repetition steps from step 5. TABLE XI. STEP 14 IN UML INFORMATION FLOW MODEL GENERATION PROCESS Generated UML Fig. 7. Full UML Information model genreated from Enterprise model of Transformation Enterprise model Information Flow Scheduled workflow for Ultrasound examination for the pet in Veterinary algorithm part element model element clinic After the implementation all the steps of transformation algorithm it can be undoubtedly declared that chosen example perfectly illustrates accuracy of the UML Information flow elements generated from Enterprise model. After 14 steps of the transformation algorithm generating of Scheduled workflow for Ultrasound examination for the pet in IV. CONCLUSIONS Veterinary clinic data from Enterprise model the 1 stage – pet In the first part of the article the Enterprise model elements owner registers his pet in veterinary clinic registration system. role variations possibilities in UML dynamic models generating Information system manages pet owner registration and process and top level of transformation algorithm are presented. services ordering, is responsible for updating information – is shown in the figure. The next part handles with detailed explanation of UML Information model transformation algorithm, which is depicted by steps. In the next part there is presented particular example, which data is stored in knowledge based Enterprise model and there are described all the stages of the example. Final part describes transformation algorithm steps for the UML Information Flow model generation from The Enterprise model and illustrates it with graphical schemes. The illustrated example shows that data stored in Enterprise model is sufficient for generating process and it is possible to claim, that every element of UML dynamic models can be generated from the Enterprise model using transformation algorithms and this can accomplish knowledge based IS development cycle design phase. Fig. 6. 1 stage of Scheduled workflow for Ultrasound examination for the pet in Veterinary clinic example is presented as UML Information model REFERENCES generated from Enterprise model. [1] Butleris, R., Lopata, A., Ambraziunas, M., Veitaitė, I., Masteika S. Full UML Information flow model after all steps of the (2015) SysML and UML models usage in knowledge based MDA process. Elektronika ir elektrotechnika. Vol 21, No 2 (2015) pp. 50-57 transformation algorithm generating Scheduled workflow for Print ISSN: 1392-1215, Online ISSN: 2029-5731 Ultrasound examination for the pet in Veterinary clinic [2] Čyras, Vytautas. (2017) Intelektualios sistemos. Vilnius: Vilniaus example is shown in the figure. universitetas, Matematikos ir informatikos fakultetas, Programų sistemų katedra. 116 p. Kenneth D. Evensen and Dr. Kathryn Anne Weiss (2010) 35 A Comparison and Evaluation of Real-Time Software Systems Modeling (5th Workshop on Applications of Knowledge-Based Technologies in Languages. AIAA Infotech@Aerospace 2010 20 - 22 April 2010, Business (AKTB 2013)). Atlanta, Georgia. [13] Lopata A., Ambraziūnas M., Gudas S., Butleris R. (2012) „The Main [3] IEEE Computer Society (2014) Guide to the Software Engineering Body Principles of Knowledge-Based Information Systems Engineering“, of Knowledge SWEBOK. Version 3.0. Paperback ISBN-13: 978-0- Electronics and Electrical Engineering, Vol. 11, No 1 (25), pp. 99-102, 7695-5166-1 ISSN 2029-5731. [4] Gudas S. Enterprise knowledge modelling: Domains and aspects. [14] Lopata A., Ambraziūnas M., Gudas S. (2012) “Knowledge Based MDA Technological and economic development of Economy. Baltic Journal Requirements Specification and Validation Technique”, Transformations on Sustainability 281–293, 2009 in Business & Economics, Vol. 11, No. 1 (25), pp. 248-261. [5] Gudas S., Architecture of Knowledge-Based Enterprise Management [15] Lopata, A., Ambraziūnas, M., Gudas, S. Knowledge-based MDA Systems: a Control View, Proceedings of the 13th world multiconference requirements specification and validation technique. Transformations in on systemics, cybernetics and informatics (WMSCI2009),) July 10 – 13, Business & Economics, 2012, 11(1(25)), 248-260. ISSN 1648-4460. 2009, Orlando, Florida, USA, Vol. III, p.161-266 ISBN -10: 1-9934272- [16] Morkevicius A., Gudas S. (2011) “Enterprise Knowledge Based 61-2 (Volume III).ISBN -13: 978-1-9934272-61-9 (Volume III) Software Requirements Elicitation”, Information Technology and [6] Gudas S., (2012) Informacijos sistemų inžinerijos teorijos pagrindai. Control, Vol. 40, No 3, pp. 181-190, 1392 – 124X Vilniaus universiteto leidykla ISBN 978-609-459-075-7 [17] OMG UML (2018) Unified Modelling Language version 2.5. Unified [7] Gudas s., Lopata A., (2007) Meta-Model Based Development Of Use Modelling// http://www.omg.org/spec/UML/2.5 Case Model For Business Function. Information Technology And [18] Perjons, E (2011) Model-Driven Process Design. Aligning Value Control, ISSN 1392 – 124X 2007, Vol.36, No.3 Networks, Enterprise Goals, Services and IT Systems. Department of [8] Küster, T., Lützenberger, M., Heßler, A., Hirsch, B. (2012) Integrating Computer and Systems Sciences, Stockholm University. Sweden by US- process modelling into multi-agent sysmtem engineering. Multiagent and AB, Stockholm ISBN 978-91-7447-249-3 Grid Systems: [19] Veitaitė I., Lopata A. (2015) Additional Knowledge Based MOF https://www.researchgate.net/publication/220535349_Integrating_proces Architecture Layer for UML Models Generation Process. 18th s_modelling_into_multi-agent_system_engineering. International Conference on Business Information Systems, BIS2015 [9] Linehana Mark H., Barkmeyerb Ed, and Hendryx Stan, (2012) The Date- International Workshop, Series: Lecture Notes in Business Information Time Vocabulary. OASIS Advancing open standards for the information Processing. society. FOIS. p. 265-378. [20] Veitaite I., Ambraziunas M., Lopata A. (2014) Enterprise Model and [10] Lopata A. Disertacija. Veiklos modeliu grindžiamas kompiuterizuotas ISO Standards Based Informations System’s Development Process. 16th funkcinių vartotojo reikalavimų specifikavimo metodas. 2004 International Conference on Business Information Systems, BIS2014 [11] Lopata, A., Veitaite, I., Gudas, S., Butleris, R. (2014) CASE Tool International Workshop, Larnaca, Cyprus, May 22-23, 2014, Series: Component – Knowledge-based Subsystem. UML Diagrams Generation Lecture Notes in Business Information Processing. Process. Transformations in Business & Economics, Vol. 13, No 2B [21] UML diagrams (2018) https://www.uml-diagrams.org/ (32B) pp. 676-696 ISSN: 1648 - 4460 [22] White, Stephen A. (2004) Process Modeling Notations and Workflow [12] Lopata, A., Veitaite, I. (2013) UML Diagrams Generation Process by Patterns. IBM Corporation. p. 23-24. Using Knowledge-Based Subsystem. Tarptautinė konferencija „15th [23] Zhou Xiaohua, Zhou Nan, Auto-generation of Class Diagram from Free- International Conference on Business Information Systems“, BIS2013 text Functional Specifications and Domain Ontology [Cite Seer X Beta], http://citeseerx.ist.psu.edu/ 36