Applications of agents in supporting an organisation'osperations addressed in academic literature largely refer to issues connected with business processes and management of an organisation knowledge, which are considered separately. In the first case, multi-agent systems are perceived as an element that performs and automates business processes of an organisation and substitutes its IT systems and users. In the second case, agent solutions are perceived as tools designed to automate or support the user in the different stages of the life cycle of the knowledge management system. The research discussed in this paper addresses the aspect of integrating both these approaches and refers to the point where both these classes of systems meet. The aim of this paper is to present the concept of an original solution ensuring integration of knowledge management systems and business process handling, which is supported by agent systems. The first part of the paper presents current research in the area of integration of software agents within business processes and the processes of knowledge processing. The second part presents the architecture of a software solution designed to support the modelling of business processes and improve these processes by ensuring the process participants access to organisational knowledge. The third part shows an example of using this architecture.
Current research into agent technologies supporting business
processes and their use in the context of supporting the building
of knowledge management systems shows that agent solutions can
support both these approaches. To be successful in an
organisation, such systems should constitute synergy of
organisational, technological and human-focused initiatives and
tools [
As figure 1 shows, knowledge management systems should support the identification of knowledge needs, gather and store such knowledge and support its sharing within business processes. Additionally, they should support creation of knowledge during performance of processes.
Such a definition of tasks results from the fact that knowledgebased organisations use intellectual capital to manufacture products and provide services, consciously manage intellectual capital and are capable of learning. In other words, knowledgebased organisations are those that adapt their offer and way of acting to knowledge gained by reflecting on how they previously acted and that consciously manage knowledge resources they possess.
In order to use software agents in organisations, it is necessary to
look for tools that will support the designing of business
processes taking place in organisations on the one hand, and will
support the performance of such processes by providing relevant
knowledge on the other hand. In particular, such solutions should
support specific stages of the life cycle of the knowledge
management system [
Of special importance in the process of building knowledge
management systems is the aspect of knowledge sharing among
users and the context of its use. Such knowledge should be
contextual, dependent on the process in which it is used and
specific tasks it supports. For that reason, it is necessary to look at
knowledge management systems from the perspective of Virtual
Community (VO), in which users share possessed knowledge in a
virtual community [
The aim of this paper is to present the concept of an original solution that ensures integration of knowledge management systems with business processes handling and is supported by agent systems.
The first part of the paper presents current research in the area of integration of software agents within the processes of knowledge processing. In particular, it indicates requirements for a solution designed to support business processes and manage knowledge about such processes. The second part presents the architecture of a software solution designed to support the modelling of business processes and improve these processes by ensuring their participants access to organisational knowledge. The third part shows an example of using this architecture.
This paper will address the issue of creating agent-oriented solutions that enable integration of knowledge management systems in the area of business processes performed by an organisation.
Figure 2 shows that the subject of using software agents within an organisation requires considering three aspects. The use of agents as an element supporting business processes, as an element supporting knowledge management and the use of agents to support the integration of both these approaches, which is the subject of this paper.
Modelling of business processes is focused on achieving a
specific goal by performing defined tasks. Users’ goal and
constitute a key element defining a business process [3] and are
often mentioned in its definition [
In the second approach, the use of software agents consists in
providing the user with appropriate knowledge required in the
context of the goal, process or task in which he/she participates
[
Tasks of software agents as part of knowledge management systems vary. They refer to specific stages of the life cycle of the knowledge management system. In the context of direct support for business processes, the task of such solutions is to ensure appropriate knowledge to the user or agent acting on his/her behalf in a specific business process. Since the research presented in this paper refers to integration of agent-supported knowledge management systems in business processes, it is necessary to analyse these solution more closely. In his work V. Dignum [4] presented a threefold division of agent solutions in the area of an organisation knowledge management (table 1).
An entity operating in isolation from other software agents, lack of mechanisms for communication with agents, orientation towards communication with a human being, possession of a local knowledge database and an artificial intelligence mechanism. Possible connection with an organisation's IT systems through tasks
with other agents (usually an interface), it uses certain communication standards and queries generation languages, uses other agents' knowledge database, e.g. interface, is able to adapt other agents' knowledge for its own needs. It operates locally in terms of cooperation with the user. It uses agents of one type. The tasks they perform and their abilities are comparable.
knowledge databases. They are created for the purpose of performing various tasks in the community in which they reside. Their task is to cooperate to achieve defined goals.
The application of the first and second types of agents can be
examined here in the context of using interface agents and
chatterbots as an element supporting the tasks performed by the
user and providing him with appropriate knowledge. The use of
such solutions supporting knowledge management in an
organisation largely results from the human nature and problems
with human memory which cause the knowledge on certain
processes to be lost and forgotten if such processes are not
regularly repeated [
The use of multi-agent solutions in the third approach largely
refers to using them to support selected actions within an
organisation. In this approach, multi-agent systems are treated as
an element that supports knowledge processing in knowledge
management systems and as an element of the architecture of such
a system. Such solutions are implemented in an organisation, for
instance, to support software management processes [
On the other hand, academic literature includes works addressing
the aspect of knowledge processing, and [
In the area of integration of knowledge management systems and
business processes, one can indicate such solutions as [
The concepts mentioned earlier show diversity of both the approaches in modelling the functionality of agent systems, resulting from the specificity of their application. In the context of agent solutions supporting business processes from the perspective of an organisation, as presented in point 2.1, an employee or agent participating in a business process should possess appropriate information and knowledge that will allow him/it to properly perform the task in which he/it participates. Additionally, such knowledge should relate to the place and time of performing a business process. From the perspective of knowledge management systems, the task of software agents is to provide the participant with appropriate knowledge, which is required in the context of the process in which he participates and the task he performs.
In this case, agent supported solutions integrating KM and BPM should have the following functions:
Extending currently used standards for describing business processes to include sources of knowledge that supports the performance of users' tasks (in the context of the process, place and time).
Enabling direct integration of organisational knowledge within any business processes taking place in an organisation within the scope of the process in which this knowledge should be used and the task that it supports.
Automating processes of assessing the functioning of knowledge management systems in terms of their usefulness in supporting business processes.
Generating new organisational knowledge at the interface of business processes and knowledge management.
Using semantic mechanisms for knowledge description for easier integration of possessed knowledge with internal organisational knowledge.
Independent operation from used IT solutions and enabling integration of any knowledge management systems and a process-oriented solution.
The software solution presented in the next chapter fulfils all the above-mentioned requirements.
This chapter will present the functional scope and architecture of a solution being built. The system is built based on JAVA language and uses various libraries, including the JADE multiagent platform, JENA mechanisms for ontology semantic processing, OWL language for knowledge representation and BPMN standard for business process modeling. The subsequent chapter will present an example of using the architecture for modelling a business process and defining knowledge resources used in it.
The aim of the solution developed is to support processes of describing knowledge resources of organisations as part of business processes taking place in them. The solution is agent oriented, as it is possible to define software agents and multiagent systems as knowledge sources for a process. The system itself also has its own multi-agent system, which is used in the process of analysing gathered knowledge and its codification by means of semantic mechanisms. This paper will present a fragment of its functionality in the area of modelling business processes in BPMN along with the adopted artefacts extending its functionality and the use of a mechanism for evaluating interface agents (chatterbot) in performing the user's tasks.
The research into the cooperation of knowledge management systems, systems supporting business processes of organisations and agent technologies resulted in the development of a system architecture that consists of seven layers (figure 3).
personalization and communication
knowledge storing Layer 1 refers to the participant of a business process. It can be a company's employee, customer, department of an organisation or other software agent. It was assumed in the architecture that the main reason for such an entity's willingness to participate in processes of an organisation is acquisition of certain knowledge. This however does not result from willingness to learn, but necessity to perform specific tasks that have been assigned to it. In this case, the knowledge it has to possess is determined by the task in which it participates, and the system of knowledge management should provide it with specific tools to support the performance of this task, taking into account the context of the process in which it participates.
Layer 2 of the system is knowledge portal and interface agents. The task of this layer is to promote knowledge on processes taking place in an organisation and to use semantic mechanisms in the form of chatterbots. It has been assumed in the architecture developed that both these elements can constitute a part of the system for an organisation knowledge management or exist outside IT systems of the organisation. For that reason, the solution proposed should support the integration of them both. This layer can be treated as the interface of an organisation knowledge management systems indicated in 2.2.
Layer 3 refers to business processes. A frequent problem encountered by users is to locate the knowledge required in the processes in which they participate, therefore this layer should support modelling of business processes. Based on analysis of notations currently used for specification of business process, the author used BPMN, extended by additional artefacts to allow to identify how agents, knowledge portals and other participants impact a process. An advantage of this approach is the fact that knowledge management systems are linked with an organisation's business processes. Figure 4 shows an example of a business process modelled in BPMN in an IT system supporting business processes of an organisation and its representation in the tool developed.
The architecture of a business process modelled in this way makes it easier to indicate which information should be made available to users while they perform actions in which they participate. The main artefacts extending the functionality of BPMN as indicated in the figure above have been presented in table 2.
Interface agent - designates possible applications of an interface agent as an element supporting the performance of a specific task by the user. It allows the user to go into the mode of evaluation of an agent's usefulness and use it to support the user's actions.
A multi-agent system - designates a possible application of a multi-agent system to substitute the user or prepare a specific knowledge resource that will be necessary in the decision making process.
Knowledge resource - designates a specific knowledge resource in knowledge portal or the Internet that can be indicated to the user. It can be a document, web service, URL identifier.
Consultant - designates a specific person who has the relevant knowledge about the performance of this task.
The process of selecting a person to perform specific
tasks has been presented in paper [
Evaluation screen
P
a
r
a
m
e
t
e
r
s
a
n
d
l
o
g
As shown in point 2.2, interface agents, in particular chatterbots,
can be an element of dissemination of the organisational
knowledge. However, their use in the context of knowledge
sharing should be subject to evaluation, due to changeability of
business processes and possible obsolescence of an agent's
knowledge. The solution proposed enables analysis of any
chatterbot agent published in the Internet. During evaluation of an
agent, the user is informed on an ongoing basis about, among
other things, which task the agent should support, how long
conversation lasts and how many questions the user has already
asked. Then, the user is asked to complete a satisfaction survey
concerning the conversation. Based on that, the system calculates
the agent's usability indicators for a specific task in which it was
used [
Layer 5 of the tool developed is the layer of a multi-agent system. In the architecture designed, it is perceived as a tool for evaluating knowledge acquired in the process of using the method indicated above and as a tool to build a multi-agent system, substituting the user in accordance with the developed methodology focused on business processes. For implementation of this layer, JADE platform was chosen and a set of agents supporting the process of knowledge evaluation and codification was prepared (figure 7).
Layer 6 of the architecture being developed is the ontology layer. It is used as an element of semantic specification of the terms used by agents and enables description of knowledge resources in an organisation. The elements of the ontology are presented in figure 8.
The use of this layer contributes to propagation of possessed organisational knowledge and is consistent with the concept of WEB 3.0. It also contributes to propagation of knowledge about an organisation and processes taking place in it, and use of this knowledge by a multi-agent system in the process of analysing organisational knowledge.
The last layer 7 is the most technical and refers to the applied system for database management which enables integration of data from any source and its processing.
The system architecture developed is designed to support business processes, therefore its central module is the module of designing business processes in BPMN notation. Such a process may come from other IT systems existing in an organisation or may be prepared based on a established model. The example shows a process of obtaining a planning permission. When applying for a planning permission in Poland, it is necessary to collect a range of documents. This process is presented in figure 9. While the process of collecting documents itself is widely known, such documents and their content may vary across different places in Poland where they are prepared. Also places of their storage will vary. Therefore, the process illustrated in figure x, though correct for a specific person, is general and for reference only. The application of the solution proposed makes it possible to extend this process by the context of its use including place and time and knowledge resources that can be used, which is presented in figure 10. As a result, the person performing such a process may become equipped with additional knowledge on its course and the source of this knowledge. He/she also has access to relevant documents that have to be completed.
In the example above, it has been diagnosed that in the city of Gdynia there is an interface agent published on the website http://www.gdynia.pl/, whose task is to support users’ action. sBy means of the interface created, it is possible to evaluate it in the context of place, time and process. As a result, the user getting access to the process may decide based on the agent's usefulness indicators whe ther to use it or not. Based on the interface of evaluating an agent's knowledge(figure 11), the multi-agent system prepares evaluation of its usefulness in this task, which is included as an element of description of a business process.
The algorithm of the process of analysing the agent's usefulness is as follows (figure 12).
BEGIN // connect OWL instances for each LOCALISATION(l) for each PROCESS(x) for each TASK(y) for each AGENT(a) if hasTask(x,y) AND hasAgent(y,a) AND hasLocalisation(y,l)
GET x and ADD to set A end if end for end for end for end for // set agent performance in connected process,task and localisation for each A(x) for each EVALUATION(e) if hasEvaluation(x,e) i:=i+1; p:=p+e.performance; x.performance := p/i end if end for end for END Based on that, the system enables calculation of the agent's usability in the context of the task that it supports, place and time of its performance (figure 13). The user may also use other knowledge resources that have been assigned to a given process.
And assign new knowledge resources to the process (figure 14), Figure 4 The process of assigning knowledge to a process by the user.
As a result, for a given business process, it is possible to develop extended BPMN process model connected to knowledge resources and a map of agents' effectiveness in supporting the different business tasks.
This paper addresses the issue of modelling agent-supported systems designed to integrate knowledge management systems within business processes performed by an organisation and its environment.
The solution proposed enables integration of various agent solutions within modelled business processes regardless of the level of agent socialisation indicated by V. Dignym The approach proposed herein offers the following advantages:
It directly relates to currently used standards for modelling business processes.
It proposes extension of BPMN to include additional artefacts connected with description of knowledge resources. The solution developed can be used as an element of a process-oriented IT system without interference into its architecture.
It allows knowledge resources to be assigned to processes, tasks, place and time of their performance. This enables better adjustment of knowledge resources to the context of knowledge use.
It allows knowledge resources to be evaluated. The current version of the solution enables evaluation of an agent interface usability and conversation with the employee. It automates the process of knowledge evaluation using a prepared multi-agent system.
It enables integration of represented knowledge resources through the semantic layer (OWL ontology) with other solutions.
Current research into this architecture focuses on the use of multiagent systems as an element supporting business processes. The original methodology of designing such a solution and elements of a tool supporting the modelling of such solutions will be the subject of the author' further research.
The project was financed from the funds of National Science Centre 2011/03/D/HS4/00782.
[1] Benotti, L.,Martínez, M. C., Schapachni"kE, ngaFg.ing High School Students Using Chatbots" in: Proceedings of the 2014 conference on Innovation & technology in computer science education, ACM 2014, pp. 63-68 [2] Das S., Shuster K., Wu C.“ACQUIRE: agen-bt ased complex query and information retrieval engine,” in Proceedings the first international joint conference on Autonomous agents and multiagent systems: part 2, 2002, pp. 631–638 of [3] De la Vara Gonzalez J. L., Diaz J. S. "Business processdriven requirements engineering: A goalbased approach" In Proceedings of the 8th Workshop on Business Process
Modeling, 2007 [4] Dignum V. „An overview of agents in knowledge management” Proceeding INAP'05 Proceedings of the 16th international conference on Applications of Declarative Programming and Knowledge Management, 2006, s. 175189 [5] Di Nitto E., Pianciamore M., Selvini P., “The role of agents in knowledge management,” in WOA, 200–23,4. pp. 29 Design,