2010 IEEE 6th World Congress on Services

A model for business process automation in service oriented systems with knowledge management technologies
Ana aa
Faculty of Computer and Information Science University of Ljubljana Traka 25, 1000 Ljubljana ana.sasa@fri.uni-lj.si

Abstract Due to increasingly demanding requirements for business flexibility and agility, automation of end-to-end industrial processes has become an important topic. Business process execution needs to support automated tasks execution as well as human tasks. The topic of this doctoral work is to enable a higher degree of automation of business processes in service-oriented systems. The doctoral work shows that for certain types of human tasks it is relevant to consider their further automation. A service-oriented architectural framework for human task execution is proposed. It improves execution of human tasks by automating and semi-automating decision making based on ontologies and agent technology. Keywords: business process automation, service-oriented architecture, multi-criteria decision making, ontologies, multiagent system .

I. INTRODUCTION Due to increasing requirements for efficiency, effectiveness and flexibility of business systems, automation of business processes has become an important topic. In the last years, the most successful and predominant approach to business process automation has become the service-oriented architecture approach. In a service-oriented architecture a business process is composed of services, which represent different tasks that have to be performed in a business system. Typically, a business process is implemented with one of the business process execution languages. Among them the prevalent language has become WS-BPEL (Web Services Business Process Execution Language). Business process execution systems perform orchestration of different services. Services can represent automated tasks or enable integration of human-performed tasks (human tasks). Integration of human tasks into business process execution is covered by a pair of specifications: BPEL4People and WSHumanTask. The topic of this doctoral work is to enable a higher degree of automation of business processes in serviceoriented systems. We focus on human tasks, because they represent those tasks of a business process, which are not automated. We define an architectural model supported by innovative methods, which extend the possibilities of automation of human tasks. Every architectural component is responsible for automation of a certain part of a human task.
978-0-7695-4129-7/10 $26.00 2010 IEEE DOI 10.1109/SERVICES.2010.71 136

In cases where automation is not enabled by the proposed model, the human task is delegated to the responsible person. For implementation of the model, we have defined an extension to the WS-HumanTask specifications. The extension allows backward compatibility with the existing specifications. In the doctoral work we did not try to propose a completely different approach to implementing human tasks as part of the business process execution, but rather enhance the existing ones with higher level support for human tasks with the objective of automating their execution as much as possible. Due to the advantages of SOA, we focus on human task improvement in SOA based systems. Furthermore, the model that we propose allows system implementation that is compliant with BPEL4People and Web Services Human Task specifications, and allows incorporation of business policies and rules into execution of human tasks. Our approach can also be applied to other types of human task systems, for example traditional workflow systems. II. RELATED WORK AND OBJECTIVES In the proposed model, ontologies are used as a means of enabling human task support and automation. Lai defines the ontology as a means of enabling communication and knowledge sharing by capturing a shared understanding of terms that can be used both by humans and by programs [6]. There are several languages available for ontology representation, such as DAML, CGs, OIL, DAML+OIL, and OWL. The OWL (Web Ontology Language) [5] is a W3Cs Semantic Web Activity effort and is expected to play an important role in helping automated processes to access information [8]. Even though our approach is based on OWL due to its ability to represent a useful group of ontology features, a high level of support, and its XML foundations, which make it appropriate to be used in conjunction with other Web technologies [5]. More specifically, our approach is based on OWLs description logics (DL) based sublanguage (OWL DL), enhanced with SWRL (Semantic Web Rule Language). SWRL is based on a combination of OWL DL and OWL Lite with the Unary/Binary Datalog RuleML sublanguages of the Rule Markup Language [7]. In the doctoral work the concept of ontology-based organisational memory (organisational ontology) is used as the basic component. It combines ontologies with decision support, more specifically multi-attribute decision making.

Multi-attribute decision making has been a topic of research for many decades; however, our work is based on multiattribute decision modelling as presented in (Bohanec and Rajkovic, 1999) [9]. The concept of organisational memory refers to stored information from an organizations history that can be brought to bear on present decisions (Walsh and Ungson, 1991) and supports sharing and reuse of individual and corporate knowledge and lessons learned (Abecker et al., 1998). Considerable research effort has been dedicated to this field of science; examples are (Walsh and Ungson, 1991 [10]; Abecker et al., 1998 [11]; Abecker and Decker, 1999 [12]; Huang et al., 2005 [13]; Dieng et al., 1999 [14]). However, there seem to be only few authors concerned with application of these approaches to the business process execution domain, similar to the proposal in this paper. Abecker et al. developed a model of a multi-layer organizational memory system, in which the upper level comprised modelling and execution of processes and tasks [11]. Kaathoven et al. systematically investigated the integration of organizational memory systems with workflow management systems [15]. Based on this work, Reimer et al. proposed a model with the objective of assisting office-clerks in doing their job, offering them active support and guidance [16]. The main difference between these papers and the presented doctoral work is that our focus is not on support of human actors when performing tasks, but on further automation possibilities of their work with the methods applied to the organizational ontology. Fox et al. [17] and Ba et al. [18] are concerned with automation in the domain of decision making based on organizational memory. The important differences in comparison with our approach are that we place the decision making in the business process context based on the SOA principles, that we focus on tacit knowledge and its translation into explicit knowledge, and that we propose a method which enables automatic querying and query answering, and thus a higher level of process execution automation. Furthermore, we do not try to propose an approach to integration of information from heterogeneous sources, which would be then used to support human actors at their work. Rather, we presume the use of the SOA integration platforms in order to gather information, which we then map to the organizational memory. In the proposed model, different human task owners are represented by agents. The objective is to enable support for human tasks with appropriate coordination characteristics. A multi-agent system (MAS) is a system composed of cooperative or competitive agents that interact with one another in order to achieve individual or common goals [19]. A wide range of proposals for business process automation improvement based on agent technologies have been discussed, however, they approach it in a different way than we do. Most authors in this field are concerned with different proposals to administer or improve process composition with agents or multi-agent systems, for example Ali et al. [20], Xu et al. [21], Blake and Gomaa [22], Jennings et al. [23]. Taveter and Wagner have proposed the Radical AgentOriented Process (RAP) based on Agent-Object-Relationship (AOR) modelling, and the RAP/AOR methodology geared

towards business process modelling, simulation and automation [24]. Their approach is agent-oriented and not service-oriented. Other authors propose to implement Web services with agent technology in order to realize complex interaction and coordination of services, for example Yi et al. [25] and Zhang and Jiang [26]. On the other hand, some authors discuss managing the organizational memory with multi-agent systems, for example Gandon et al. [27] and Abecker et al. [28]. Even though the agents in our system work as intermediaries between the business processes, human actors and the organizational ontology, their main role is to provide coordination support for human task execution. Therefore, our work is not closely related to this specific field. III. OVERVIEW OF THE MODEL AND DESCRIPTION OF
CONCEPTS

The model is based on observation that business process and human task execution can be seen as an opportunity to capture the knowledge based on which the execution is performed, and to capture the knowledge which is created during the execution. The goal is to enable direct reuse of the knowledge and possible automation for succeeding human task executions. In order to enable this, we define a decision making method, which is based on multi-criteria decision making and develops a decision model in an OWL ontology [2]. The model enables the use of this method in human task execution as part of business process execution. With this it enables a higher level of automation of human tasks, which are based on decisions. The model also enables a higher level of automation of human tasks, which require collaboration between different human actors. We propose an innovative approach to defining business collaboration protocols, which are the basis for collaboration between agents in a multi-agent system, which represent the human actors. The approach enables separation of protocol definition and implementation and, with this, loose coupling of protocols and implementation. This means that changes in a collaboration protocol do not require changes in the implementation of the multi-agent system or the implementation of the behaviour of the agents, which use the protocols in their acts. Figure 1 gives an architecture overview of the proposed model. The model comprises two main services: service for human task execution (SIUO) and organizational ontology service (SOO). The figure illustrates different layers which indicate a wider context for using the services of the model. The business process layer demonstrated the use of services in business process execution. The human task execution layer comprises the SIUO, user interface and people performing the tasks. The SIUO service is responsible for execution of human tasks. Based on the type of a human task it enables different levels of human task automation. It is responsible for execution of elementary and collaboration human tasks. In case of a basic pattern human task, it triggers the corresponding pattern based process. A pattern based process recursively triggers SIUO for elementary human task execution. The central component of SIUO is a multiagent system. When the service is triggered the broker agent

137

accepts the task. In case it is a basic pattern human task, the agent passes the request to the human task pattern manager which triggers the corresponding process. Otherwise, the agent determines the task owners. If it is an elementary human task, the broker agent delegates it to one of the elementary human task execution agents. If it is a collaboration task, it delegates the corresponding subtasks to different elementary human task execution agents that

represent the corresponding business actors. The SIUO service uses the SOO service. The ontology layer comprises the SOO and human tasks that refer to development and use of organizational ontology [3]. Besides the domain knowledge, the organizational ontology comprises decision models corresponding to different business actors and the protocol library.

Figure 1. Overview of the architecture model

As decision making and collaboration are among the most important activities of employees in business systems, the model improves a significant part of human tasks. It represents an important step closer to the vision of end-toend business process automation. It heightens the degree of automation of human tasks and consequently enables more efficient business process execution. Another important advantage is a higher level of decision traceability. IV. CONCLUSION

automation of human tasks and, by this, enabling a higher degree of automation of business processes.

REFERENCES
[1] A: aa: A model for business process automation in service oriented systems with knowledge management technologies. Doctoral Thesis, Faculty of Computer and Information Science, University of Ljubljana, Slovenia, 2009. A. Sasa, M.B. Juric, and M. Krisper, Service-Oriented Framework for Human Task Support and Automation, IEEE Transactions on Industrial Informatics , vol. 4, 2008, pp. 292-302. A. Sasa and M. Krisper, Knowledge Management in ServiceOriented Systems, Information modelling and knowledge bases

The doctoral work represents an integral solution to improving efficiency of human tasks within the business process context in service-oriented systems and defines a set of new and innovative methods, which are based on ontologies and multi-agent systems. Their common goal is

[2]

[3]

138

[4]

XXI, (Frontiers in artificial intelligence and applications, vol. 206). Amsterdam [etc.]: IOS Press, cop. 2010, pp. 89-104. OASIS, "Web Services Business Process Execution Language Version 2.0," OASIS Standard, http://docs.oasisopen.org/wsbpel/2.0/wsbpel-v2.0.html, Apr. 2007. W3C, OWL 2 Web Ontology Language Structural Specification and Functional-Style Syntax, http://www.w3.org/TR/owl2-syntax/, 2009. L. F. Lai, "A knowledge engineering approach to knowledge management," Information Sciences , Vol. 177, No. 19, Oct. 2007, pp. 4072-4094.

[5] [6]

W3C, "SWRL: A Semantic Web Rule Language, Combining OWL and RuleML," http://www.w3.org/Submission/SWRL/, May 2004. [8] F. Baader, D.L. McGuiness, and D. Nardi (Eds.), Description Logic Handbook: Theory, Implementation and Applications , Cambridge University Press, Cambridge, UK, 2003, ch. I. Horrocks, D.L. McGuiness, and C. Welty, "Digital Libraries and Web-Based Information Systems." [9] M. Bohanec and V. Rajkovi, "Multi-Attribute Decision Modeling: Industrial Applications of DEX," Informatica , Vol. 23, No. 4, Oct. 1999, pp. 487-491. [10] J.P. Walsh and G.R. Ungson, "Organizational Memory," The Academy of Management Review, Vol. 16, No. 1, Jan. 1991, pp. 5791. [11] A. Abecker, A. Bernardi, K. Hinkelmann, O. Khn, and M. Sintek, "Toward a technology for organizational memories," IEEE Intelligent Systems, Vol. 13, No. 3, 1998, pp. 4048. [12] A. Abecker and S. Decker, "Organizational Memory: Knowledge Acquisition, Integration, and Retrieval Issues," In Proc. XPS , Wrzburg, Germany, 1999, pp.113-124. [13] C.-C. Huang, T.-L. Tseng, and A. Kusiak, "XML-Based Modeling of Corporate Memory," IEEE Transactions on Systems, Man, and Cybernetics, Vol. 35, No. 5, Sep. 2005, pp. 629 640. [14] R. Dieng, O. Corby, A. Giboin, and M. Ribire, "Methods and tools for corporate knowledge management," International Journal of Human Computer Studies , Vol. 51, No. 3, Sep. 1999, pp. 567-598. [15] R. van Kaathoven, M. A. Jeusfeld, M. Staudt, and U. Reimer, "Organizational Memory Supported Workflow Management," In Proc. 4. Int. Tagung Wirtschaftsinformatik , A.-W. Scheer, M. Nuttgens (Eds.), Electronic Business Engineering, Physica-Verlag, Heidelberg, Germany, 1999, pp. 543-563. [16] U. Reimer, A. Magelish, and M. Staudt, "EULE: A Knowledge-Based System to Support Business Processes," Knowledge-Based Systems, Vol. 13, No. 5, Oct. 2000, pp. 235-331.

[7]

[17] M. Prietula, K. Carley, and L. Gasser (Eds), Simulating Organizations: Computational Models of Institutions and Groups , AAAI/MIT Press, Menlo Park, CA, 1998, ch. M.S. Fox, M. Barbuceanu, M. Gruninger, and J. Lin, "An Organisation Ontology for Enterprise Modeling." [18] S. Ba, K.R. Lang, and A.B. Whinston, "Compositional enterprise modeling and decision support," Information Systems and E-Business Management , Vol. 6, No. 2, Mar. 2008, pp. 137-160. [19] B. Henderson-Sellers and P. Giorgini (Eds), Agent-oriented Methodologies, Idea Group Inc., Hershey, PA, 2005, ch. P. Giorgini and B. Henderson-Sellers, "Agent-Oriented Methodologies: An Introduction." [20] S. Ali, B. Soh, and T. Torabi, "A Novel Approach Toward Integration of Rules Into Business Process Using An Agent-Oriented Framework," IEEE Transactions on Industrial Informatics, Vol. 2, No. 3, Aug. 2006, pp. 145-154. [21] Q. Xu, R. Qiu, and F. Xu, "Agent-Based Workflow Coordination for Supply Chain Management," Transactions of Nanjing University of Aeronautics & Astronautics , Vol. 20, No. 1, Jun. 2003, pp. 112-117. [22] M.B. Blake and H. Gomaa, "Agent-oriented compositional approaches to services-based cross-organizational workflow," Decision Support Systems , Vol. 40, No. 1, Jul. 2005, pp. 31-50. [23] N.R. Jennings, T.J. Norman, and P. Faratin, "Autonomous Agents for Business Process Management," International Journal of Applied Artificial Intelligence, Vol. 14, No. 2, 2000, pp. 145-189. [24] B. Henderson-Sellers and P. Giorgini (Eds), Agent-oriented Methodologies, Idea Group Inc., Hershey, PA, 2005, ch. K. Taveter and G. Wagner, "Towards Radical Agent-Oriented Software Engineering Process Based on AOR Modelling." [25] Y. Li, K-M Chao, M. Younas, Y. Huang, and X. Lu, "Modeling emarketplaces with multi-agents Web services," In Proc. 11th Int. Conf. on Parallel and Distributed Systems , Fukuoka, Japan, 2005, pp. 175-181. [26] T.I. Zhang and H. Jiang, "A Framework of Incorporating Software Agents into SOA," In Proc. Artificial Intelligence and Soft Computing (ASC 2005), Benidorm, Spain, 2005. [27] F. Gandon, A. Poggi, G. Rimassa, and P. Turci, "Multi-Agents Corporate Memory Management System," Journal of Applied Artificial Intelligence , Vol. 16, No. 9&10, Oct.-Dec. 2002, pp. 699720. [28] A. Abecker, A. Bernardi, and L. van Elst, "Agent Technology For Distributed Organizational Memories: The Frodo Project," In Proc. 5th International Conference On Enterprise Information Systems , Angers, France, 2003, pp. 3-10.

139