AGENT-BASED SYSTEMS DESIGN FOR VIRTUAL

ORGANISATIONS FORMATION

Tiemei Irene Zhang

School of Management and Information Systems, Faculty of Business and Law, Victoria University

Footscray Park Campus, Australia

Keywords: Agent-based systems, Methodology of analysis and design, Virtual organisation.

Abstract: These days, organisations must adapt to business and technical changes which are vital in a competitive and

ever-changing business environment. To meet the dynamically changing requirements, virtual organisation

is widely recognised as an effective solution. Multiple agent technology has been actively discussed and

recognised as the merit of flexibility and adaptability. Thus, this technology can be applied to virtual

organisations. To prevent from being impeded by using incongruous approaches to designing agent-based

systems, systematic approach is essential to incorporate a variety of organisation business units that are

required to meet current and future needs. This paper aims to present a systematic approach including

process of analysis and design to designing virtual organisations. It also demonstrates a case study for

application of this approach.

1 INTRODUCTION

These days, organisations must adapt to business

and technical changes which are vital in a

competitive and ever-changing business

environment. However business processes in

organisations are so complicated that traditional

organisations cannot efficiently cope with this

challenge. Furthermore, the organisations that are

geographically dispersed face more difficulties. To

meet the dynamically changing requirements (Patel

et al, 2005), strategic alliance of business processes

is necessary to be put into practice (Schneider,

2009). Virtual organisation is widely recognised as

an effective solution to be able to address this issue.

As a virtual organisation is the goal-driven dynamic

assembly of organizational tasks (Shan et al, 2006)

and is created for specific business need, it can

simplify business processes.

To establish virtual organisations, IT

infrastructure is fundamental and critical to provide

feasible and supportive solutions. Efficiency and

effectiveness of the infrastructure determines the

success of a virtual organisation in goal sharing,

collaborations, ownerships and functionalities.

Multiple agent technology has been actively

discussed and it is recognised as the merit of

flexibility and adaptability (Kollingbaum et al,

2006), (Udupi & Singh, 2006). Its application to

virtual organisations is a wise decision. To prevent

from being impeded by using incongruous

approaches to designing agent based systems,

systematic approach to designing virtual

organization is essential to incorporate a variety of

organisation business units that are required to meet

current and future needs. Although some of

researchers have applied multiple agent systems to

virtual organisation formation (Leong et al, 2006),

how to systematically design a virtual organization

is neither explicit nor well-developed.

This paper aims to present an agent-based

systematic approach to designing virtual

organisations. Following introduction and

background sections, §3 depicts how a virtual

organisation is logically structured. §4 describes an

engineering process to analyse and design agents for

a virtual organization. A case study is then

presented to demonstrate application of this research

in §5. Finally, §6 concludes the paper and discusses

further research.

Irene Zhang T. (2010).

AGENT-BASED SYSTEMS DESIGN FOR VIRTUAL ORGANISATIONS FORMATION.

In Proceedings of the 2nd International Conference on Agents and Artiﬁcial Intelligence - Agents, pages 54-61

DOI: 10.5220/0002728900540061

 SciTePress

2 BACKGROUND

2.1 Virtual Organisation

A virtual organization is described as: “a temporary

network of independent companies – suppliers,

customers, even erstwhile rivals – linked by

information technology to share skills, costs and

access to one another’s markets. It will have neither

central office nor an organization chart.” (Byrne,

1993) The term “virtual organization” is also used

to cover a wide range of geographically dispersed

organizations. These may be linked by common

goals with each requiring a new style of

management and reassessment of the role of

management. The Internet and related technologies

are used as a means of communication and

collaboration (Collins, 2002).

In reality, the components that make up a virtual

organization – individual employees, teams,

departments, units or firms – are geographically

distributed, functionally or culturally diverse,

electronically linked and connected via lateral

relationships. These attributes enable the

organization to dynamically modify business

processes to meet market demands, to coordinate via

formal and informal contracts, to define the

boundaries of the firm differently over time or for

different customers or constituencies, and to re-

arrange relationships among components as needed.

(Pedersen & Nagengast, 2008)

2.2 Multiple Agent Systems in Virtual

Organisation

Researchers have actively discussed and addressed

issues of applying software agents to virtual

organisations. After investigating 14 AOSEs

(Agent-Oriented Software Engineering)

methodologies, Leong et al (Leong et al, 2006)

claimed that these methodologies do not have strong

contributions in terms of virtual organisation (VO)

formation and management. VO formation and

management still remain as open research problems.

In addition, they also pointed out that application of

these technologies is lack of software engineering

techniques.

Shan et al (Shan et al, 2006) proposed an agent-

mediated service framework in which an ad hoc

approach is used to design their agents. To improve

the adaptive ability of the agent design methods in

uncertain environment, Zhang (Zhang, 2009)

proposed a modelling method based on role theory

and Agent UML. Pournaras et al (Pournaras et al,

2009) developed an approach to making use of

virtual organisation consisting of agents for

distributed management of resource utilisation.

In summary, it is widely recognised that virtual

organisations are characterised by flexibility and

adaptability. Software agent and multiple agent

systems are actively investigated and determined to

be an efficient way for virtual organisation

formation and management. Although some

methodologies have been developed for agent-based

systems’ analysis and design, due to their

limitations, systematic methodologies and software

engineering techniques for virtual organisation

formation are still open for investigation.

3 VIRTUAL ORGANISATION

ARCHITECTURE

The layered architecture pattern (Buschmann et al,

1996) has been widely accepted as a standard in

network design and software engineering. This

architecture reduces the coupling and complexity of

dependencies that may occur in individual problems.

The architecture can be used to structure

applications that can be decomposed into groups of

subtasks. Each group of subtasks must also be at a

particular level of abstraction.

To apply this architectural pattern to an agent-

based system, the following forces need to be

considered:

 An agent system is complex and spans several

levels of abstraction;

 There are dependencies between neighbouring

levels, with two-way information flow;

 The software architecture must encompass all

aspects of agency;

 The architecture must be able to address simple

and sophisticated agent behaviour.

The following several reasons discuss why

agents in a virtual organisation should be

decomposed into layers:

 The agents of higher levels, or with more

sophisticated behaviour, depend on lower

level capabilities

 Layers only depend on their neighbours, and

there is two way information flow between

neighbouring layers for communication.

 The agents of lower levels focus on specific

tasks and services.

 The layers can be identified from modelling of

the agents’ real world

AGENT-BASED SYSTEMS DESIGN FOR VIRTUAL ORGANISATIONS FORMATION

Each layer provides a level of abstraction and

certain services to the layer above it, while hiding

the implementation of its services from the higher

layers. Also, each layer passes both data and control

information to its corresponding neighbours. This is

similar to communication within an organisation in

the real world. This communication is classified

into categories described as below (Ivancevich et al

1997):

 Downward-communication: communication

flows downward from higher levels of the

hierarchy to lower levels. This type of

communication serves some organizational

functions, such as job instruction, job

rationale, and procedures.

 Upward-communication: communication flows

from a lower level in an organisation to a

higher level. This type of communication

provides organizational functions, such as

feedback of a problem, and information

required for decision making.

 Horizontal-communication: communicators are

in the same level. This is necessary for the

coordination and integration of diverse

organizational functions to get social need

satisfaction.

Therefore, a layered architecture is proposed for

agent-based virtual organisation as shown in Figure

Interface Layer (IL)

Security Layer (AL)

Service Broker Layer (SBL)

Service Layer (SL)

Organizational Layer (OL)

Ontology Layer (ONL)

Services

Organisational

Database

Data

files

Figure 1: Architecture of agent-based virtual organisation.

In this figure, there are six different layers that

make the system more compartmentalised and

modularised. Four of our layers are hierarchical,

and dependencies occur only between adjacent

layers. However both Security and Ontology layers

are utilised by all of the other layers. In the case of

the hierarchical layers, the complexity of the lower

layers is hidden from the higher levels and from the

user. This layered architecture provides a

mechanism allowing adjustment and configuration

of business processes according to organisations’

needs and facilitates integration and reuse. The

proposed agent-based virtual organisation promotes

the advantages, such as flexibility, adaptability,

scalability and security.

An organization requires the ability to accurately

identify the user who is making requests. The

process that verifies and records the user’s identity is

called authentication. This process is designed to

employ an access-control-list that contains a single

entry authorized to grant capabilities for other

layers. In actuality, agents in every layer in an

agent-based virtual organisation have to get security

clearance from the security layer before they can

request services.

Ontology Layer collectively maintains a

knowledge base of the different terminology and

concepts that are employed over the whole

organization. This layer is similar to Security Layer

because ontology can be shared by agents in several

other layers. A given ontology describes and

specifies the terminology or language that would be

used for specifying requests for information. Thus

Ontology Layer is used to adapt or translate a

request to other modules in different layers.

Interface Layer is tied closely to an individual

human’s goals, and it can be used to predict the

user’s intentions and to request services provided by

the remaining modules. The agents in this layer act

on behalf of users to relay specifications and obtain

results

Service Broker Layer predicts or models the

intentions of the overall organization and then

provides services to users via Interface Layer. Both

Interface and Service Broker Layers have their own

objectives, and they may compete with each other to

achieve their goals. Negotiation may occur between

Interface and Service Broker Layers. In order to

provide better services to users, Service Broker

Layer should process the information obtained from

other layers and send results to the user via Interface

Layer.

Service Layer is used to provide services on

behalf of the organization. This layer differs from

Organizational Layer that controls resources. It

represents and provides the high level services of the

organization. These services can be formed by

encoding expertise and by utilizing Organisational

Layer. Service Layer should have policies that

identify if it can provide services to certain

individuals.

ICAART 2010 - 2nd International Conference on Agents and Artificial Intelligence

Organizational Layer can be used to manage the

organizational resources. One of the important tasks

in this layer is to gather data from various sources

such as data banks, advanced product-planning data,

individual cost estimates, supplier documentation

and engineering test & field data. Data can be

managed by either the legacy applications or new

services provided from cloud computing. It can be

recorded in different ways and must be adjusted to

be consistent and comparable (Fabrycky &

Blanchard, 1991). An individual agent in

Organizational Layer is closely tied to sources of

data.

4 AGENT DESIGN PROCESS

To design agents for a virtual organisation, it is

critical to smoothly and efficiently transform

requirements to implementation through analysis

and design. Doing so requires systematic approach

in SDLC (software development life cycle). The

following figure describes an approach for software

agent analysis and design:

Goals

Goal Cases and

Beliefs

Roles

RRC

(Role-Responsibility-

Collaborator)

Develop Goal Cases

GCB

(Goal Cases,

Collaborators, and

Beliefs)

RGC

(Responsibility-Goal-

Collaborator)

AgentsCompose Roles as Agent

Assign Goals to Responsibility

Figure 2: Systematic approach to designing agents.

In this figure, both Goals and Roles can be

identified from use cases. A goal is an objective or a

desired state that can be achieved or ascertained by

an agent. A goal identifies what is to be done, and it

should change less often than more detailed

processes/activities. This is because a process or

activity identifies how things are to be done. Goals

are important to agent-based systems because agents

are autonomous and proactive. Goals can be

achieved through their autonomous and proactive

behaviour which is defined as a collection of

scenarios (goal cases) about agents’ interactions.

Each scenario describes a sequence of plans to

handle events that the agent initiates. An agent can

start a goal case when the corresponding goal is

triggered. The use of goal cases also helps with

traceability because they are developed according to

goals that are derived from the system requirements.

Roles can be identified from use case and

relevant role models (Kendall, 1999). A role model

contains a set of roles, their corresponding

responsibilities and collaborators. A role can be

described by role model name, role types,

responsibilities, and collaborators, which is

represented as a RRC (Role-Responsibility-

Collaborator)

card described below:

Table 1: RRC card.

Role model name

Role type Responsibility Collaborator

Names of

Roles

List all

responsibilities

List all

collaborators

Examining all role models, roles can be

categorized and composed into many sub-groups

according to the description of the layers in Figure 1.

Each role or composed role can be described as a

RGC (Responsibility, Goal, and Collaborator) card

as shown in Table 2.

Table 2: RGC card.

Role

Responsibilities Goals Collaborators

List responsibilities List goals List collaborators

In this card, role name, its associated

responsibilities and collaborators are documented

and goals are assigned to the appropriate

responsibilities.

An agent plays at least one role and each role

possesses its goals and collaborators. Subsequently

goal cases and beliefs with these goals are

determined. As a result of this, an agent can be

specified as GCB (Goals, Goal Cases, Collaborators,

and Beliefs) card described below:

Table 3: Agent specification template (GCB card).

Agent Name:

Goal Goal Case Collaborator Belief

List all

goals

List all goal

cases

List all

collaborators

List all

beliefs

AGENT-BASED SYSTEMS DESIGN FOR VIRTUAL ORGANISATIONS FORMATION

5 CASE STUDY

To demonstrate application of the systematic

approach developed in section 4, a case study has

been developed for designing a virtual organisation,

which is able to gather information for life cycle cost

estimation. The following subsections present the

case, the analysis & design process and

implementation of the agent-based virtual

organisation.

5.1 Overview

This case study aims to establish a virtual

organisation to gather information from various data

sources and then estimate life cycle operation and

supporting (O&S) costs for a product such as a

computer system as shown in Figure 3:



Computer

System

Computer Case

Power Supply

Electronic Board

Keyboard

Floppy Disk Drive

Floppy Disk Interface

Hard Disk Drive

Hard D isk Interface

Monitor

Monitor Case

Screen

M onitor Control BoxCD ROM

Printer

Printer Interface

Modem System

M odem Interface

Modem

Modem Cable

CD ROM Interface

Figure 3: A computer system breakdown structure.

In this figure, the computer system is broken

down into subsystems, which are comprised of

assemblies (subsystems) and parts. Each assembly

may in turn be comprised of other assemblies and

parts. Furthermore, these assemblies can recursively

be divided until no more assembly is found. The

agent-based system goes through all parts and

subsystems to accumulate the different costs in

certain cost categories. Moreover, the CASA model

(Manary, 1996) that provides algorithms to assess

life cycle cost is chosen as a life cycle costing model

for this case study.

5.2 Role Identification and

Composition

To illustrate how to identify roles, consider an agent

(in Service Layer) which is responsible for cost

estimation by cooperating with other agents to

gather information. This agent should play a service

role to provide expertise that the organization

possesses to external customers and internal units.

The main purpose of the service role is to distribute

work to the other roles and then to compute a final

result by using the returns from the others. This

behaviour can be modelled by the Master/Slave role

pattern where the service role plays the role of

Master (Aridor & Lange, 1998; Buschmann et al,

1996). In addition, the service role can receive

requests from other roles; hence it also needs to play

the role of Slave.

This agent also plays roles to collaborate with

other layers, such as Security Layer, Service Broker

Layer and Ontology Layer. With Security Layer, it

plays Client and Subject roles in the Bodyguard role

model; with Service Broker Layer, it plays Subject-

proxy role in the Broker role model; with Ontology

Layer, it plays Client and Target roles in the Adapter

role model. These roles can be composed to form

the agent. As each role associates its responsibilities

identified by relevant role model, it is not difficult to

produce the following RRC card:

Table 4: RRC card of Service Role.

Service Role

Role type Responsibilities Collaborators

aster

(Maste

Slave)

1. To collaborate with

the Slave and send

requests

Organizational

Layer

Subject-proxy

(Broker)

. Register the service

3. Provide the service

Service

Broker Layer

lient &

arget

(Adapter)

. To send message to

Adapter and

collaborate with

the Target

5. To receive the

message sent by

Client

6. To perform a task

and send a reply

>Ontology

Layer

lient &

Subject

(Bodyguard)

. To request the

permission of a

service

8. To accept the

notification of a

service

>Security

Layer

Note:

>represents servers of the collaborating agent.

<represents clients of the collaborating agent.

ICAART 2010 - 2nd International Conference on Agents and Artificial Intelligence

In a similar manner, role models, such as

Observer, Broker, Master/Slave, Manager/Owner

Bodyguard and Adaptor can be applied to Figure 1

to identify potential roles for all layers in a virtual

organisation. These roles can be composed as

agents and instantiated for this case study. All

instantiated agents can be structured as shown in

Figure 4.

Figure 4: Organization of agents.

In this figure, the agents at the lowest level, such

as Project Manager Agent, Support Agent, Inventory

Agent and Resource Agent are called Organizational

Agents. They are specific to their own information

sources. The legacy databases, represented by

cylinders, can be accessed by only Organizational

Agents. In Service Layer, there are two agents,

Planning Agent and Estimation Agent that are

categorised to Service Agents. These agents provide

services based on the information that

Organizational Agents provide. Security Agent in

Security Layer collaborates with User Agent and

Maintenance Agent in our application. User Agent

in Interface Layer acts on behalf of a user to

communicate with Maintenance Agent in Service

Broker Layer. Information Agent in Ontology Layer

is used to access the ontology database that stores

ontological commitments between different

organizations. Both Service Agents and

Organizational Agents interact with it.

To develop a RGC card, examine Estimation

Agent in more details. Estimation Agent aims to

estimate the life cycle cost. Once receiving the

request to estimate O&S cost, it has to achieve goals

such as “Estimate Labour”, “Estimate Equipment”,

“Estimate Material”, “Estimate Management”,

“Estimate Support”, and “Estimate Miscellaneous”,

one by one. To achieve these goals Estimation

Agent has to ask the individual agent in

Organisational Layer for information which the

organisational agent is specific with. Table 5

documents the typical RGC card of this agent.

Table 5: RGC card of Estimation Agent.

Estimator

Responsibilities Goals Collaborators

o collaborate

ith the Slave

nd send requests

•

Request Labour

Item Cost

•

Request

Equipment Item

Cost

Resource Agent

•

Request

Material Item

Cost

Inventory

Agent

•

Request

Management

Item Cost

Support Agent

•

Request Support

Item Cost

•

Request

Miscellaneous

Item Cost

egister

•

Estimate O&S

cost

<Maintenance

Agent

rovide the

service

•

Estimate Labour

Cost

• Estimate

Equipment Cost

• Estimate

Material Cost

• Estimate

Management

Cost

• Estimate

Support Cost

• Estimate

Miscellaneous

Cost

>Planning Agent

o send message

o Adapter and

ollaborate with

he Target

•

Alternative

>Information

Agent

Security

Agent

Supplier’s

Products

Support

Agent

Resource

Agent

Inventory

Agent

Planning

Agent

Estimation

Agent

Maintenance

Agent

User Agent

Projects

Project

Manager Agent

Specifications

Information

Agent

Ontology

Interface Agent

Broker Agent

Service Agent

Organizational

Agents

Data

Source

Support

Information

AGENT-BASED SYSTEMS DESIGN FOR VIRTUAL ORGANISATIONS FORMATION

5.3 Goals, Goal Cases and Agent

Specification

Using GCB card to specify an agent requires

providing the agent’s goals and goal cases. To

demonstrate how to develop an agent specification,

consider material cost estimation as an example. To

estimate the total cost of sub-category of material,

the goal “Estimate Material Cost” should be

achieved. This high level goal can be broken down

further to the sub-goals, e.g. “Search Alternatives”

and “Request Material Item Cost” which is used to

calculate the material cost for a part or a subsystem

in the product. To achieve a goal, a goal case needs

to be developed. Table 6 shows a goal case to

achieve the goal “Estimate Material Cost”.

Table 6: Example of goal case.

Estimate Material goal case (GC 6.8.1)

Pre-condition:

Start date, end date, maintenance level, and part id.

Flow of events

Basic paths:

1. The goal case starts when Estimation Agent

attempts to achieve the “Estimate Material

Cost” goal.

2. Estimation Agent asks Information Agent to find

keywords for information required in the

organization (GCI.1.0).

3. Estimation Agent requests Planning Agent for

factors used in the algorithms of CASA model.

4. Estimation Agent asks Inventory Agent for those

factors using the “Request Material Item Cost”

goal case (GC6.8.2).

5. Estimation Agent calculates item cost by using

the CASA algorithms.

6. Estimation Agent stores the cost in the agent

database.

Alternatives:

In step 2, 3, 4, if failure to achieve the goal,

Estimation Agent asks the user to enter it.

Post-condition

Estimation Agent stores material item cost data

From this goal case, all material item costs are

determined and they can be accumulated and

determined to be a total material cost stored as the

belief “Material Cost”. Other goal cases for costs of

labour, equipments, management and support in this

research can be processed in a similar manner.

These costs will be feed to the CASA model for the

assessment of total O&S cost and life cycle cost.

As a result of Tables 4, 5 and 6, the following

table documents relationships among all goals, goal

cases, collaborators and beliefs for Estimation

Agent:

Table 7: Estimation Agent specification (GCB).

Estimation Agent specification

Goal Goal

Cases

Collaborators

(Agents)

Beliefs

equest

abour Item

ost

C A.4.1*

Resource

esource

ost

aterial

ost

anage-

ent Cost

Support

ost

iscellan

ous Cost

equest

quipment

tem cost

C A.4.2*

equest

aterial Item

ost

C 6.8.2*

Inventory

equest

anagement

tem Cost

C A.4.4*

Support

equest

Support Item

ost

C A.4.6*

equest

iscellaneous

tem Cost

C A.4.8*

stimate TOS

C 6.8.3*

Maintenance

stimate

abour cost

C A.4.3*

Planning

stimate

quipment

ost

C A.4.0*

stimate

aterial cost

C 6.8.1

stimate

anagement

ost

C A.4.5*

stimate

Support Cost

C A.4.7*

stimate

iscellaneous

ost

C A.4.9*

lternative

CI.1.0*

Information

Note:

*represents goal cases developed in this research but

they are not shown in this paper. They are included

only for completeness.

5.4 Implementation

The JACK agents (Busetta et al, 1999) are used to

implementing the proposed agents in the virtual

organisation for life cycle cost estimation. To

construct JACK agents, four class-level constructs:

Agent, Database, Event, and Plan are used. The

Agent construct declares types of events and plans

used to handle the events. It does not only have

methods and data members just like objects, but also

contains database relations that an agent can use to

ICAART 2010 - 2nd International Conference on Agents and Artificial Intelligence

store beliefs, descriptions of events that the agent

can handle, and plans that agent use to handle the

events. Table 8 shows a map between GCB card

and JACK constructs.

Table 8: GCB card to JACK agent.

GCB card Goal Goal

Case

Collaborator Belief

JACK agent

Constructs

Event Plan

Agent Database

The mapping rules above are used to implement

the designed agents in the case study. The similar

implementation has been done by Zhang (Zhang,

2002).

6 CONCLUSIONS/FUTURE

WORK

This paper has presented a systematic approach to

design software agents for virtual organization

formation. In this approach, role pattern and layered

architecture are used to simplify agent analysis and

design processes. Various card templates used to

define agent behaviours are developed to assist agent

design. The case study has been used to verify and

validate the approach. Subsequently, the approach

can be applied to different businesses which involve

data and services from various sources. As a result

of this research, the approach has demonstrated its

advantages of flexibility, adaptability, scalability and

robust. Future work will be focused on extension of

the case study to incorporate services provided by a

cloud computing platform.

REFERENCES

Aridor, Y., and Lange, D.B., 1998. Agent Design Patterns:

Elements of Agent Application Design. Autonomous

Agents (Agents’98), Minneapolis, pp. 108-115.

Busetta, P., Ronnquist, R., Hodgson, A. and Lucas, A,

1999. Light-Weight Intelligent Software Agents in

Simulation. SimTech 99, Melbourne Australia, 1999.

Bryne, J.A. 1993. The Virtual Corporation. In Business

Week, February 8, p.103.

Buschmann, F., R. Meunier, H. Rohnert, P. Sommerlad,

and M. Stal, 1996. Patter-Oriented Software

Architecture: A System of Patterns, Wiley, USA.

Collins, P., 2002. Virtual and networked organizations,

Capstone Publishing, Oxford.

Pournaras, E., Warnier, M. and Brazier, F., 2009. A

Distributed Agent-based Approach to Stabilization of

Global Resource Utilization. In 3

International

Conference on Complex, Intelligent and Software

Intensive Systems, pp.185-192. Fukuoka Institute of

Technology (FIT), Japan, March, 16th – 19th 2009.

Fabrycky, W.J., and B. Blanchard, 1991. Life-Cycle Cost

and Economic Analysis, Prentice-Hall, Inc., New

Jersey, USA.

Gary P. Shneider, 2009. Electronic Commerce, Course

Techonology, Cengage Learning, ISBN-13: 978-1-

4239-0305-5

Ivancevich, J., Olekalns, M. and M. Matteson, 1997.

Organisational Behaviour and Management,

McGraw-Hill Book Company Australia Pty Ltd.,

Richard D. Irwin, Inc.

Kendall, E. A., 1999. Role models, aspect oriented

programming and agent engineering. Technical report,

British Telecom,

Kollingbaum, M., Mehangjiev, N. and Brown, K., 2006.

Engineering Organisation-Oriented Software. In

WISER '06. 2

International Workshop on Workshop

on interdisciplinary Software Engineering Research,

pp23-28. Shanghai, China, May 20 - 20, ACM Press,

New York.

Leong, P., Miao, C. and Lee, B., 2006. Agent Oriented

Software Engineering for Grid Computing. In

CGRIDW’06. Sixth IEEE International symposium on

Cluster Computing and Grid Workshops, pp-2.

Singapore, May 16-19.

Manary, J.M., 1996. DSMC’s CASA Model Still Going

Strong. PM: JANUARY-FEBRUARY 1996, p37-40.

Patel, J. etc, 2005. Agent-based Virtual Organisation for

the Grid. In AAMAS’05. 4th International Joint

Conference on Autonomous Agents and Multi-agent

Systems, Utrecht, Netherlands, July 25-29.

Pedersen, C. and Nagengast, J., 2008. The virtues of the

virtual organization. Strategic HR Review, 7 (3). pp.

19-25. ISSN 1475-4398

Shan, B., Han, Y., and Sun, W., 2006. An Agent-Mediated

Service Framework Facilitating Virtual Organizations.

In Proceedings of the 10

International Conference on

Computer Supported Cooperative Work in Design.

2006, LNCS 4402, pp. 438-446, 2007.

Udupi, Y. B. and Singh, M. P., 2006. Multiagent Policy

Architecture for Virtual Business Organisations. In

SCC’06, 3

IEEE International Conference on

Services Computing, pp. 44-51. Chicago, USA,

September 18 – 22.

Zhang, J., 2009. Research on Modeling Method of Virtual

Enterprise in Uncertain Environments. In ICEE’09,

2nd International Conference on Computer

Engineering and Technology, pp. 541-544. Singapore,

January 22 - 27.

Zhang, T. I, 2002. Agent-based Information Gathering

System for Life Cycle Costing, Ph.D. thesis, RMIT

university, Melbourne, Australia, 2002.

AGENT-BASED SYSTEMS DESIGN FOR VIRTUAL ORGANISATIONS FORMATION