Various attempts have been undertaken to further automate the BPM lifecycle. One of the recent initiatives in this area is the Semantic Business Process Management approach as pursued within the SUPER project. It aims at bridging the gap between business and IT worlds by increasing the degree of automation within the BPM lifecycle using Semantic Web and Semantic Web services technologies. In order to ful l the SBPM vision, enterprises as well as their environment need to be properly described. The main contribution of this paper is a set of ontologies for describing organizational structures and examples showing how they may be combined with further organizational information, e.g., business functions and business roles, to support the automatic analysis of business processes.
A business process may be de ned as a set of related, ordered activities that contribute to the production of good(s) or the delivery of some service. It emphasises how the work is done within an organization and by its organization members. Therefore, organizations have their own context for any running business process even if they operate in the same domain. This context embraces information like used resources, strategies, enterprise structure as well as roles and functions. However, when modelling business processes using current standardized notations (e.g. BPMN) a lot of information, especially on organizational aspects, cannot be represented.
Various attempts were undertaken to achieve automation of the BPM lifecycle. One of the most advanced initiatives in this area is the concept of Semantic Business Process Management developed within the SUPER project (Semantic Utilised for Process Management with and between Enterprises). It aims at bridging the gap between business and IT worlds by increasing the degree of automation within the BPM lifecycle using Semantic Web and Semantic Web services technologies.
In order to ful l the SBPM vision, apart from the semantic description of the
process ow (control structure), the process content description is also required.
The process content relates to the enterprise and its environment and therefore,
must rely on a proper organization description. Furthermore, for these aspects
to take part in any automated processing they need to be expressed in a formal
and machine readable form. The goal of this paper is to present a set of
ontologies that support capturing organisational information as required for realizing
the Semantic Business Process Management vision. We present four ontologies
supporting the semantic representation of a process content and position them
within the entire organizational framework developed within the SUPER project.
The overall approach we propose in based on the use of ontological descriptions
of process ow enriched with additional descriptions of the related resources, the
organisational entities involved, the roles required, etc as proposed earlier in [
The remainder of the paper is organized as follows: rst, we discuss the scope for the semantic description of an organization. Then, four ontologies, namely Organizational Structure Ontology, Organisational Units Ontology, Business Functions Ontology and Business Roles Ontology are shortly described followed by an example of process description with the use of developed ontologies. Finally, we position our work within the state of the art in the area of semantic representation of an enterprise. Finally we present some conclusion and introduce lines for future research. 2
In order to describe an organization for the needs of the Semantic Business Process Management, an appropriate vocabulary needs to be provided. It should allow for the description of both processes and process artefacts (process content). As mentioned, the volume of knowledge needed to adequately describe all organizational details relevant to any given element of a business process is immense. In order to represent this knowledge, relevant contextual information need to be extracted as well as important elements and aspects need to be identi ed.
The main aspects that need to be captured are the process structure and
enterprise domain speci c terms (see Figure 2). The process related vocabulary
is important to ensure a proper and unambiguous representation of control ow
of processes within and between enterprises, while the latter part of stack is to
provide a speci c representation of a domain of given organization. Therefore,
the process-related information (i.e. process ontologies) describes the structure of
a process (i.e. tasks, control structures, links etc.), whereas organisation related
ontologies provide a description of enterprise artefacts (i.e. actors, functions etc.)
that are utilised or involved in the process realization. The domain ontologies
provide the additional information speci c to an organization from a given
domain. As already stated in the introduction, the focus of this paper is on the
organizational ontologies that provide a high-level view on the organization and
process-related space. Therefore, it is to provide vocabulary and constraints for
describing the environment in which processes are carried out from the
organization's perspective i.e. to allow describing actors, roles and other resources that
may be referred to in the process de nitions [
Domationlogies n O
O r g
Ontoloangiiesastional Process Ontologies
Task 1
Task 2
Task 3
Within the SUPER project we identi ed the following areas of the organisation description: business functions, business roles, organisational units, organisational structure, process resources, enterprise strategy and modelling guidelines. Each area is addressed by the respective ontology. All ontologies form an ontology stack for the needs of organizational aspects representation and usage within the business process modelling phase. The approach applied in SUPER was built based on previous achievements in the eld of organisation modelling and was shaped by an extensive BPM practice while the preceding approaches involved mainly traditional managerial concepts. Hence, the SUPER view is more compact (e.g. it decomposes the institutional internal structure into functions and roles) whereas it also seems to be less human-centred (comparing e.g. skills and sta elements to resources of a process). The important similarities embrace the more general view on the organization; highlighting the elements that in uence the institutional environment (i.e. strategy and modelling guidelines) in a comprehensive manner.
De ning a fully comprehensive model of an enterprise, is a particularly challenging and demanding task which would require an extremely extensive and complex conceptualisation which it would be di cult to manage, adapt, and extend. Therefore, we propose instead a modular ontology stack that provides basic notions, systematizes them and thus provide the foundation for further development able to address the needs of a particular enterprise.
Below a short description of ontologies constituting the SUPER organisational ontologies stack follows: { Business Functions - providing vocabulary to describe the hierarchy of different functions that may be carried out within a company (e.g. marketing, nance, HR) in order to enable vendor and domain independent classi cation of company processes and process fragments providing abstraction over single tasks constituting processes. { Business Roles - representing roles in the organisation e.g. Designer, Process
Modeller, IT Expert, CEO. { Organizational Structure and Organisational Units Ontologies presenting detailed view on the organisational structure of an entity being described. { Process Resources - describing applications and resources that should be spent when carrying out processes (e.g. documents, IT systems, or any other resources needed to accomplish a task) or that may be a result of a certain task in a process. { Enterprise Strategy - in order to model information related to an enterprise strategy, targeted markets, bene ciaries of the company and some surroundings that in uence the processes carried out in a company. { Modelling Guidelines - used to describe policies and rules that should be taken into account during processes modelling.
In the following sections, the Business Organizational Structure Ontologies as well as the Functions and Business Roles Ontologies constituting a part of the ontologies stack are presented. The nal shape (concepts and relations) of ontologies is a result of thorough analysis of the most representative ERP systems together with the critical approach towards the related work described in section 7. The expert knowledge was also an important source of inspiration. All mentioned ontologies were modelled using the WSML formalism. 3
The Organizational Structure Ontology (OSO) focuses on supporting the definition of organisational structures. Following literature from this domain, the organisational structure may be de ned as a structure and/or hierarchy of an organization and how its constituents work together to achieve common goals. Therefore, an organizational structure encompasses: departments, employees, their responsibilities, resources etc. as well as relations among them. The Organizational Structure Ontology is designed as an upper level ontology and therefore provides a main structure and relations aiming at achieving the domain independency. OSO structure is a result of analysis and mixture of a few described in a literature organisational structure models. The developed structure was based and/or inspired by [2{7].
The structure of the developed ontology makes it easy for other ontologies created within SUPER to import OSO concepts and use them as super concepts (e.g. Business Roles Ontology, Business Functions Ontology, Resource Ontology, and nally Organizational Unit Ontology may import OSO). OSO links all organizational ontologies developed so far in the SUPER project and enables description of organizational structure.
The following concepts are included in the OSO ontology:
{ Organization - a social arrangement which pursues collective goals, which
controls its own performance, and which has a boundary separating it from
its environment Legal Entity - an entity that can enter into a legal contract
[
This ontology is designed to provide a common vocabulary for description of "any
recognized association of people in the context of the enterprise. In a hierarchical
structure, it may be the corporation, a division, a department, a group or a
team. In addition, it may be a committee, a task force, a project management
organization, a class (for education) and so on"[
However, a variety of existing departments as well as naming standards,
forced us to make some simpli cation. The developed ontology includes only the
most common organizational departments. We assumed that we are modelling
Organizational Units Ontology for a production enterprise (therefore, in order
to describe some highly speci c organizational units of e.g. publishing house or
security company, appropriate concepts should be de ned). The Figure 4 depicts
the idea of OUO, whereby the arrows in the gure represent the is-a relation.
As we decided to use Organizational Unit de nition proposed by OMG in [
Business Functions and Business Roles Ontology Within this section we give a short overview of Business Functions and Business Roles Ontology.
For a more comprehensive description please refer to [
In the functional approach, the structure of the business organization is being (partially) expressed with the functions carried out by an abstract entity (a group, an agent or a department). BFO has been created in order to provide a skeleton of business functions that would be common for every enterprise, regardless of the domain the enterprise operates in. The BFO ontology is a complete construction, nonetheless it is supposed to be a comprehensive - yet generic - starting point for further development. Therefore, BFO includes only the high level view on the functional side of an enterprise that should be extended by domain-speci c ontologies.
In turn, the BRO ontology introduces the vocabulary needed to describe roles of both internal and external actors as performers of process' tasks. Therefore, the purpose of the Business Roles Ontology (BRO) is to specify the common meaning of concepts related to roles acquired and played by actors within the organizational microenvironment. We de ne a BusinessRole as a set of expected behaviours, prerogatives and obligations featured by an actor. Any actor may play more than one role and these roles may change depending on the situation (context). Three main elements of the BRO knowledge model encompass: InternalRole, ExternalRole and InternalRoleType concepts. Although BRO is a generic, top-level ontology, it aims at the ful lment of the mentioned constraints and provides a minimal, yet comprehensive, set of concepts. By a microenvironment we mean both the inside structure of an organization or an enterprise and its close surroundings. The ontology is designed in a way that supports its further development and extensions. 6
This section presents a simple business process described using organizational ontologies. We decided to keep the example very simple to assure clarity intelligibility of presented idea.
The process was modelled using Business Process Modelling Notation (BPMN)
serialised to sBPMN (ontological version of the BPMN speci cation developed
within the SUPER project [
The example presents an internal process of ordering printing paper in the Human Resources Department. It includes the following steps: 1. An employee goes to an o ce assistant and informs her about the shortage of printing paper. 2. The o ce assistant contacts a warehouse manager and places the appropriate order. 3. The warehouse manager makes the reservation for the paper in the system. 4. The delivery man delivers the paper to the o ce.
To describe the process, respective elements of the company must be described. This includes the description of the actors involved (i.e. o ce assistant, warehouse manager and delivery man), the functions performed by them as well as organisational units implicated. Let's name the o ce assistant MrsA. She works on position of an o ce manager. As an o ce manager, her roles in the company are: managing human resources and managing o ce. The following listing presents de nition of appropriate instances of ontology concepts in WSML.
instance ManagingOfficeRole memberOf bro#OrganizationalAndAdministeringRole instance HumanResourcesManagingRole memberOf bro#HRMRole instance OfficeManager memberOf oso#OrganizotionalPosition playsRole hasValue {HumanResourcesManagingRole, ManagingOfficeRole} assignedTo hasValue {ouo#HRDepartment} instance MrsA memberOf oso#Person
worksAs hasValue {OfficeManager} MrB works as a warehouse manager in the Inventory Department, thus his only role in the company is managing the warehouse.
instance ManagingWarehouseRole memberOf bro#ResourceAdministeringRole instance WarehouseManager memberOf oso#OrganizationalPosition playsRole hasValue {ManagingWarehouseRole} assignedTo hasValue {ouo#InventoryDepartment} instance MrB memberOf oso#Person
worksAs hasValue {WarehauseManager} MrC works as a delivery man in the Logistics Department, his role is distributing supplies.
instance SupplyDistributingRole memberOf bro#SupplementalRole instance DeliveryMan memberOf oso#OrganizationalPosition playsRole hasValue {SupplyDistributingRole} assignedTo hasValue {ouo#LogisticsDepartment} instance MrC memberOf oso#Person
worksAs hasValue {DeliveryMan} Finally, the process of processing the order for supporting commodity needs to be created.
instance SupportingCommoditySupplyOrderProcesing memberOf sbpmn#Process isEnclosedInBusinessFunction hasValue bfo#SupplyOrderProcessing In Figure 6 our exemplary process is presented. As comments, fragments of WSML code showing the use of OSO, OUO, BRO and BFO ontologies are included. Such a semantic annotation of the presented process allows for categorization of the process model (and its fragments) as well as a clear and informed assignment of roles/tasks/responsibilities. Moreover, as the entire process content is described in the machine accessible manner, the automatic processing and reasoning is possible. Therefore, employing appropriate algorithms developed within the SUPER project e.g. translation from the BPMN to BPEL process representation is possible. Moreover, business analysts are provided with new facilities for querying the process space may this be pre or post-execution. It is also possible to reason on who is responsible for what, who is doing what, what are the roles/functions associated with speci c tasks. In addition, it is possible to answer questions such as i) which organizational units are responsible for carrying out given functions; ii) which organizational units are participating in a given process; and iii) which processes will be a ected by changes in business process given (and how these changes will in uence people, resources, etc.)?
Answers to such exemplary questions may be of great importance in various situations. Such information is needed whenever the organizational structure of an enterprise is being managed and amended e.g., when the top management reinstance NotificationOfDemand memberOf sbpmn#ManualTask ... hasPerformer hasValue MrsA ...
instance PaperReservation memberOf sbpmn#ManualTask ... hasPerformer hasValue MrB ...
instance PaperDelivery memberOf sbpmn#ManualTask ... hasPerformer hasValue MrC ...
Notification of demand
Paper reservation
Paper delivery structures the functional areas of the enterprise by outsourcing certain functions. Enhancing business process models with semantic descriptions enables reasoning on the process execution results, the main functionalities o ered, process categorization, organizational units committed etc. Furthermore, a clear representation of the departments and/or employees interactions and their commitment in enacted processes allows for constant monitoring of their workload. Such an approach can noticeably improve the quality of business process management in the company. In case of BRO the sociological side of business organization is the clue to understand the potential use of this model. Normally, the sociological knowledge (meaning the behaviour, expectations, obligations, common interests, agent grouping con gurations) is an ordinary and widespread knowledge when it comes to human beings. Yet, because of lack of experience and very impaired possibilities of real-life interactions, in case of automated information processing and machine reasoning there is a strong need of modelling such knowledge explicitly. Without it, no mechanism should be expected to know what is a customer, how a supervisor di ers from any other employee or what does a stakeholder concept stand for in the context of a given organization. 7
Within this section a short overview of the related work is provided. A detailed
survey of research done in the eld of formal enterprise and organization
description may be found in[
In 1982 William E. McCarthy modelled the REA (Resource-Event-Actor)
ontology [
As a result of works conducted in the TOVE project [
The e3-value ontology provides modelling constructs for representing and
analyzing a network of enterprises exchanging items of economic value with each
other [
Business Process Management Ontology, described in [
A novel approach is presented in [
As far as the organizational structure is concerned interesting models were described in [2{4]. Each of them represents more or less the same level of detail, the main di erence being the described relations and the scope of the models. The resulting structure of OSO ontology presented in this paper is a based on the analysis and combination of solutions taken from each of them.
The abovementioned initiatives provided an inspiration and foundation for developing the organizational ontologies. The structure of REA and ContextBased Enterprise Ontology are probably closest to our vision, yet they do not separate vocabulary from business process structure, which has been one of our main goals. 8
This paper deals with the semantic annotation of business processes in the context of the so-called Semantic Business Process Management approach. In particular, it focuses on the Organizational Structure Ontology divided into upper level organizational structure ontology and organizational units ontology. Additionally, two other ontologies, Business Functions Ontology and Business Roles Ontology are shortly discussed in order to show how the organizational structure ontology ts into the stack of organizational ontologies. The structure of the ontologies is inspired by the earlier work in the eld, and extended in order to support the level of automation pursued. The solution proposed in this article is more comprehensive, provides extended exibility when describing processes as well as adjusting the ontologies to the needs of the speci c enterprise. Moreover, it is compatible with the recent emerging standards of the Semantic Web.
We do not foresee any major constraints in the use of the presented models.
The separation of concepts used for more advanced description of organization
into subontologies is a proper solution for achieving the desired power of
expressiveness. Nonetheless, the proposed ontologies should be further validated
and possibly extended in order to meet the real-life challenges including more
detailed levels of information . In this respect, future work will be devoted to
the extension of the ontologies and their integration with other models focussed
on supporting Business Process Analysis [