E-GOVERNMENT AND POLICY SIMULATION IN

INTELLIGENT VIRTUAL ENVIRONMENTS

Fotis Aisopos

, Magdalini Kardara

, Philipp Senger

, Roman Klinger

, Athanasios Papaoikonomou

Konstantinos Tserpes

, Michael Gardner

and Theodora Varvarigou

Distributed, Knowledge and Media Systems Lab, National Technical University of Athens, Athens, Greece

Department of Bioinformatics, Fraunhofer SCAI, 53754 Sankt Augustin, Germany

School of Computer Science and Electronic Engineering, University of Essex, Colchester, U.K.

Keywords: Virtual Worlds, Intelligent Environments, e-Government, Policy Simulation, Role-playing.

Abstract: Recently, there has been an increase of interest in the social impact of virtual reality technologies, as virtual

worlds experienced an increase of their popularity in various social groups (teenagers, businesses). The

growing trend for people to spend more time in suchlike virtual spaces implies a demand for intelligent

virtual environments, that could mimic the real world as a simulation and provide functionalities and tools

for behaviour analysis and adaptation to user preferences. Within the context of e-government, this paper

presents the ongoing project +Spaces, which is developing a range of virtual environment tools. The

platform architecture is presented and technical challenges involved in creating intelligent virtual spaces for

e-government as well as draft policies to be used in role-playing simulations are discussed.

1 INTRODUCTION

With the emergence of the 3D internet and the

appearance of 3D virtual worlds (VWs) and social

networks, for social and business purposes, virtual

environments have established themselves into

people's life. VWs consultancy K Zero has reported

that registered accounts in the virtual worlds sector

have reached 1.7 billion during the fourth quarter of

2011, having an increase of 125% over the last two

years (KZero Worldswide, 2012). Social networking

services, however, still have the biggest market

share, as these communities are visited by 67% of

the global online population (Nielsen Reports,

2009).

The increase of popularity implies a great social

impact of those environments over various user

groups. Thus, there is a great deal of interest in using

immersive virtual environments for a range of

“intelligent” applications, adapting to the

preferences and behaviours of the users. The user

virtual reactions as well as their social graphs

(graphs representing the connections between users

of a virtual world) would be useful to be exploited as

much as possible. Social networks and virtual

worlds harbour a huge amount of structured as well

as unstructured data, which can be beneficial for

political analysis.

Davies (Davies et al., 2008) reported on how

intelligent virtual worlds can be used to simulate real

spaces such as an intelligent campus. He described

iCampus, which allows people to inhabit a virtual

environment, interacting with each other and with

devices in a similar way to a real campus. Within

this space there were simulations of smart spaces

which included the University of Essex’ iSpace and

MiRTLE facilities (Gardner et al., 2008).

+Spaces (“Positive Spaces” – Policy Simulation

in Virtual Spaces) is exploring how virtual

environments (virtual worlds and social networks)

can be used to allow government bodies measure

public opinion in a large scale and maximize the

value from prospective policy measures by

leveraging the power of these communities. Thus, it

is building a range of intelligent applications,

varying from polling and debating applications to

more advanced role-playing simulations for such

environments. This paper presents those applications

as well as scenarios illustrating how they can be

used to achieve mass participation of citizens and

extract valuable conclusions from their behaviour.

129

Aisopos F., Kardara M., Senger P., Klinger R., Papaoikonomou A., Tserpes K., Gardner M. and Varvarigou T..

E-GOVERNMENT AND POLICY SIMULATION IN INTELLIGENT VIRTUAL ENVIRONMENTS.

DOI: 10.5220/0003935701290135

In Proceedings of the 8th International Conference on Web Information Systems and Technologies (WEBIST-2012), pages 129-135

ISBN: 978-989-8565-08-2

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

The paper is organized as follows: Section 2

describes the current practices on e-government and

presents the idea of +Spaces. Section 3 presents the

+Spaces applications as well as the supporting

intelligent mechanisms for virtual environments,

while Section 4 provides implementation scenarios

and the platform architecture. Finally, Section 5

presents initial results of the first project pilot

including the poll application and Section 6 the

conclusions of the current work.

2 PROBLEM FORMULATION

AND RELATED WORK

2.1 Current Practice

Over the last years, there is a growing trend for

citizens to influence the decision-making process,

something that yields many important benefits (Irvin

et al., 2004), also encouraged by new technologies

allowing mass participation. Government agencies

also seek to refine their policy-making processes by

assessing the impact of prospective policies on the

society prior to their implementation. The advent of

the internet and its pervasiveness in Western

societies led many developers to create applications

based on mass participation, using opinion polls and

forums to exploit the crowd sourcing philosophy

(Jaeger, 2003).

Nonetheless, these tools have not met with the

high expectations that agencies had of them (Scott,

2006 and Grönlund, 2010), for a number of reasons:

Firstly, most e-government tools are predominantly

applied using their own portals, lacking scaling

(Oostveen et al., 2004) as they focus on specific

groups without being able to easily extend them (e.g.

beyond national level). Secondly, due to their very

nature, they have a narrow focus, both in time and

subject, on specific and simple topics, failing to

provide a holistic, long-term view. Thirdly,

participation is restricted to a minority, namely

citizens motivated enough by the issue in question to

invest time in participating, or specific target groups

(Phang et al., 2008). Government agencies need to

rather take advantage of existing forums in which a

representative distribution of individuals from all

socio-economic and cultural backgrounds exists.

2.2 e-Governance in Virtual Spaces

In this context, the term “Virtual Space” (VS) is not

only restricted to 3D virtualized environments, such

as Second Life, but as well referring to other worlds

allowing users to socialize online, such as social

network platforms. However, although 2D networks

also provide the ability for their users to “live” in

these spaces, it is the 3D virtual worlds that emulate

some complex socio-economic aspect of real-life

through an open-market concept, not reaching,

though, the popularity of 2D networks, as shown

above.

The most evolved virtual worlds have rules and

regulations analogous to a legislative framework. In

worlds such as massively multi player online role-

playing games (MMORPGs) like World of Warcraft,

and life simulation games like Farmville, virtual

economies exist with their own currency and rules

defining the possession and persistence of property.

Virtual items can be bought and sold between

individuals for real money (Guo et al., 2007).

Spontaneous effects including mass-protests, crime,

and harassment, with consequent population

migrations have been witnessed in these

environments (Jenkins et al., 2007). In each of these

examples, participants tend to extend their

personality through their avatars (Castronova, 2005).

Virtual spaces in general allow individuals to

present profiles of themselves, oriented towards

work-related contexts, romantic relationship

initiation, or connecting those with shared interests

(Ellison et al., 2006). Social networking is now often

used in political campaigns (Powell et al., 2011).

The existence of the conditions mentioned above

implies an economic and political system with

properties similar to those seen in real economies.

Several aspects of economic theory could be used to

study virtual environments, as significant research

has been conducted towards the direction of relating

virtual world economics to the economics of the real

world (Ondrejka, 2004), (MacInnes, 2004). The

relative liberty and anonymity between users,

provided through the use of avatars, can reduce the

power of peer pressure to conform to societal or

cultural norms and stereotypes. For example, in

2007 Italian workers of an international organization

went on strike. About 1850 workers expressed their

strike related activities across the organization's

various sites in Second Life (Attwell, 2008).

2.3 The +Spaces Perspective

As derived from the previous paragraphs, virtual

spaces seem as an ideal “microsociety”, mimicking

the real world, where large numbers of citizens can

debate on simulated legislations and policies. Exact

real-world simulations, however, involve some form

WEBIST2012-8thInternationalConferenceonWebInformationSystemsandTechnologies

130

of implementation of a game-like environment e.g.

public transport or a marketplace, which represents a

real-world situation. This can be impossible to

implement, requiring an accurate representation of a

real-world setting within an artificial virtual

environment. The alternative scenario decided to be

followed in this work, is to focus on existing virtual

spaces in which users are already collaborating with

one another, such as online e-learning virtual worlds

mentioned above, already providing a simulation of

a real-world activity (i.e. learning).

Figure 1: Virtual demonstrations in Second Life.

+Spaces aims to provide policy makers with the

means for testing their legislation through such

applications in a number of virtual environments,

coupled with technologies for aggregating, filtering

and analysing this information. Information derived

from those environments is then represented in an

appropriate fashion to enable policy makers to draw

conclusions on the potential outcomes of the policies

they propose. Our work also targets at increasing

participation in this process through social

networking sites and extend it to new social

communities who are currently not heard. 2D virtual

environments can help e-governance applications to

achieve critical mass in the participation.

3 CREATING POLICY MAKING

APPLICATIONS IN VIRTUAL

SPACES

3.1 +Spaces Applications

+Spaces aims to act as a mediator between

government applications and virtual environments.

The functionality of +Spaces is exposed through a

usable API that allows application developers to

implement and deploy experiment applications on

various virtual spaces and provides tools to support

those, by processing and analysing aggregated data,

grouping user behaviours and protecting the

platform from malicious use. Interaction with

citizens in +Spaces involves sensitive personal data,

thus it is necessary for all users to give their consent

prior to participating in an experiment. For

validating the platform, three policy making

applications have been selected to be used, presented

in the following sections.

3.1.1 Polling

A polling application is appropriate for investigating

policy issues requiring only simple feedback. Polls

typically include presenting a topic with one or more

questions. Questions may encourage various types of

responses, selecting a single or multiple possible

answers or providing a level of agreement with a

statement, giving unstructured text-based responses.

From a virtual environment perspective, a poll is a

simple forum where people provide feedback

through their avatars/profiles. In +Spaces, it is

formed as a public polling booth that only one user

can take control of at a time, taking into account

users’ privacy. The main advantage of polls over

other applications is that they enable strictly defined

questions, so they are appropriate in cases where

such definition is possible.

3.1.2 Debating

Unlike polling, the debate application requires not

only quantitative but mainly qualitative feedback

from the citizens. Debating experiments are initiated

by presenting a topic of interest, raising various

arguments and questions, and calling for responses.

Participants' contribution to debates, synchronous or

a-synchronous, is typically not structured, and

relates not only to the initial statement but also to

other users’ contributions. Through debate

application, users of virtual environments experience

the sense of participating in a meeting, hearing

opinions, and even expressing their opinions. For

virtual worlds, a scenario as such can be illustrated

in the work of Drew Harry (Harry et al., 2008),

exploring novel ways to support meetings in those

worlds, including elements of both polls and debates

(Figure 2).

3.1.3 Role-playing Simulation

The role-playing simulation applications are the

most challenging to implement, providing a virtual

E-GOVERNMENTANDPOLICYSIMULATIONININTELLIGENTVIRTUALENVIRONMENTS

131

Figure 2: Debating in a virtual world meeting space.

space in which the participants are assigned roles

(e.g. policy maker, civil servant, local government

agent, citizen, etc.) to act out a particular

government policy scenario (e.g. implementation of

a new waste removal service by private contractors)

through an online role-play simulation activity. In

role-playing, even when acknowledged to be

artificial, participants adapt their personalities to

their roles, as in the case of Stanford prison

experiment (Zimbardo et al., 2000). The role-play

simulations can take place in a virtual world visually

recreating the location of the intended policy (e.g.

town hall, local street), using the appropriate

avatar’s outfits and respective rights, to approach

simulation as much as possible. By monitoring

users’ behaviour throughout the whole role-playing

simulation and analysing it via an intelligent data

analysis service, we can investigate the effect of the

applied changes to each role.

3.2 Intelligent Virtual Spaces

Data&Statistical Analysis

Reputation&Recommendation

IdentityManagement&Authorization

DataManagement&BaselineSecurity

Virtual

Environments

+Spaces

Policy Makers

Real World

simulation

exploitation

provision

Figure 3: +Spaces data flow.

In Figure 3, an overview of the information flow in

+Spaces can be seen. The last stage before all data is

collected from virtual environments to be distributed

to policy makers are the recommendation and

reputation functionalities, as well as the data and

statistical analysis. Those operations are adding

'intelligence' to the platform, processing data from

the virtual spaces to derive new information. They

include tasks such as performing data correlations

and graphs for policy makers, recommendations for

virtual environments users etc.

3.2.1 Recommendation Service

In +Spaces, the recommendation service uses

content-based techniques, to provide personalized

recommendations to virtual environment users,

proposing experiments of interest (Guy, et al.,

2009)and recommend participants for a new

experiment. This supports the platform by attracting

people to different polls and debates, based on their

own and their friends’ behaviour. Based on their

profile, they may be invited in a particular virtual

space, for a role-playing simulation.

Thus, this service also assists towards creating a

critical mass for the project scenarios. A

recommendation may be initiated by the experiment

owner himself, by providing an initial list of

potential participants during the creation process,

which will be passed on to the Recommendation

Service by the middleware. This capability makes

the virtual environment more attractive to users, and

helps with the diffusion of +Spaces applications

among users (Freyne et al., 2009).

3.2.2 Reputation Service

The reputation framework depends on users’ actions

and social network. For instance, the popularity of

some users or the positive responses to them in

debates can contribute to their reputation.

Information derived from this service is essential to

analyse the user behaviour, filtering the data

delivered to the data analysis service. The main

scope of this tool is to identify patterns of malicious

behaviour, and assign low reputation rates to such

users.

Additional types of reputation are involvement,

influence, etc. that weigh experiments' participants.

Based on the reputation rate is passed on to the data

analysis mechanism, the system assesses each user’s

credibility and trustworthiness. The results of the

reputation service is used in order to support the

experiments moderators, who can potentially ban or

mute malicious users.

WEBIST2012-8thInternationalConferenceonWebInformationSystemsandTechnologies

132

3.2.3 Data Analysis Service

The data analysis of the different types of

applications provides aggregated information about

the results of a question of interest. In polls,

according to a main question we get the distribution

of different answers in pie or bar charts. To further

investigate the results, another question can be

selected as a filter for the results, or a filter on

demographic data can also be applied to those.

Debates are less structured, consisting of a

chronologic and not contextual sequence of different

free text contributions. To get an overview of the

debate topics, we incorporate latent Dirichlet

allocation as topic models (Blei et al., 2003). This

results in lists of the most important words

representing a topic and links between them, in an

intuitive overview of the whole debate structure and

development over time. Additionally, an adapted

sentiment analysis (Kim et al., 2006) is applied, to

assign a measure to each topic.

In role-playing simulations, a combination of the

previous tools combined with clustering algorithms

leads to a specific and detailed analysis of each. The

outcome provides an analysis, which records the

changes in the opinion of the participants on

arguments and/or time. This can especially be

coupled to debate analysis, to understand the reasons

of certain user behaviours.

4 IMPLEMENTATION

ARCHITECTURE

To bring all these capabilities together, a service

oriented architecture is created, based on loosely

coupled distributed services, ensuring flexibility and

fault tolerance and also safeguarding the anonymity

and confidentiality of the sensitive data processed

within the platform. Towards interoperability, we

abstract each virtual environment’s functionality to a

basic set of services that are exposed in a

standardized way, to create a so-called middleware.

Hosting the services outside of the platforms enables

combinations into larger application workflows, thus

we deliver both platform and software as a service.

A high level view of the architecture is illustrated

in Figure 4. The +Spaces platform acts as the middle

layer between the underlying virtual platforms and

the e-Government application layer that allows

policy makers to deploy and manage applications

through a Front-End. A key concept is to support

existing and future virtual spaces through adaptors

on top of their existing interfaces (APIs). Seamless

access to those adaptors is provided through the VS

Management Layer, acting as an abstraction layer on

top of them. The +Spaces Middleware handles the

communication and data flow between the

government end-users, the VS Management Layer

and the built-in and external analysis services. Apart

from the three built-in analysis services mentioned,

the platform potentially supports the addition of

more services with similar or different

functionalities, to provide a wider variety of tools.

Figure 4: +Spaces architecture.

5 FIRST PILOT EVALUATION

04/26 05/04 05/12 05/20 05/29

Number of contributions

Date Month/Day

All spaces

Facebook

Twitter

Wonderland

Figure 5: Distribution of the participation of the first pilot

phase splitted into different virtual spaces compared to all

spaces.

The first pilot focused on deploying and testing the

poll application in three different Virtual Spaces

(Wonderland

, Facebook, Twitter). Thus, we

recruited participants applying one real policy in

order to measure the technical feasibility of running

Open source 3D virtual collaboration toolkit | Open Wonderland,

http:// http://openwonderland.org/

E-GOVERNMENTANDPOLICYSIMULATIONININTELLIGENTVIRTUALENVIRONMENTS

133

an experiment in several environments. A specific

poll concerning smoking banning in all public

spaces was created, monitoring and aggregating the

information back to the governmental agency.

Figure 5 provides the distribution of the user

participation for all virtual spaces. The total number

of answered questions in these 77 participations of

the first pilot is about 473 (~7 answers per poll).

Initial trials were successful, with the platform

being stable even when receiving tens of concurrent

actions and the end-users being satisfied from depth

and user friendliness of the poll results analysis.

6 CONCLUSIONS AND FUTURE

WORK

In this paper, we presented the work being

performed for +Spaces to create an intelligent

platform that will support policy makers and involve

virtual citizens in the policy making process. The

interoperability supported, along with intelligent

auxiliary services (Recommendation Service,

Reputation Service and Data Analysis Service),

illustrate the added value of the platform.

The first pilot provided valuable feedback, which

led to various correction actions concerning the user-

friendliness and the intelligent services functionality.

A more concrete evaluation along with some more

useful conclusions will be extracted during the next

two pilots operation and evaluation, including

debates and role-playing simulations respectively.

The policy will be the same for both, so as to make a

correlation between them and richer results

expected, will be helpful to evaluate data analysis

efficiency.

ACKNOWLEDGEMENTS

+Spaces is funded by the EU FP7, theme ICT-

2009.7.3: ICT for Governance and Policy

Modelling, under Contract No. 248726.

REFERENCES

KZero Worldswide. (2012), Virtual World registered

accounts reach 1.7bn in Q4 2011:

http://www.kzero.co.uk/blog/virtual-world-registered-

accounts-reach-1-7bn-q4-2011/

Social Networks and Blogs Now 4th Most Popular Online

Activity, Ahead of Personal Email, Nielsen Reports.

March 9, 2009.

M. Davies, V Callaghan and M Gardner, “Towards a

Mixed Reality Intelligent Campus” in Proceedings of

the 4

International Conference on Intelligent

Environments, IE04, 2008

M. Gardner, J. Scott and B. Horan, “Reflections on the use

of Project Wonderland as a mixed-reality environment

for teaching and learning”, ReLIVE 08 conference,

Open University, UK, 2008

R. A. Irvin and J. Stansbury, “Citizen Participation in

Decision Making: Is It Worth the Effort?”, Public

Administration Review Volume 64, Issue 1, pages 55–

65, February 2004, DOI: 10.1111/j.1540-

6210.2004.00346.x

P. Jaeger, “E-government around the world: lessons,

challenges, and future directions”, Government

Information Quarterly, Vol. 20, No. 4. (2003), pp.

389-394

J. K. Scott, “E” the People: Do U.S. Municipal

Government Web Sites Support Public Involvement?”,

Public Administration Review Volume 66, Issue 3,

pages 341–353, May - June 2006

Å. Grönlund, (2010). Ten Years of E-Government: The

‘End of History’ and New Beginning. Electronic

Government (pp. 13-24).

A. M. Oostveen and P. Besselaar, "From small scale to

large scale user participation: a case study of

participatory design in e-government systems". In

Proceedings of the eighth conference on Participatory

design: Artful integration: interweaving media,

materials and practices - Volume 1 (PDC 04), Vol. 1.

ACM, New York, NY, USA, 2004, 173-182.

DOI=10.1145/1011870.1011891

C. W. Phang and A. Kankanhalli, "A framework of ICT

exploitation for e-participation initiatives". Commun.

ACM 51, 12 (December 2008), 128-132.

DOI=10.1145/1409360

Y. Guo and S. Barnes, "Why people buy virtual items in

virtual worlds with real money". SIGMIS Database 38,

4 (October 2007), 69-76.

DOI=10.1145/1314234.1314247

P. S. Jenkins, "Virtual Worlds as a New Game Theoretic

Model for International Law: The Case of Bilateral

Investment Treaties", (2009) 6:1 SCRIPTed 4,

http://www.law.ed.ac.uk/ahrc/script-ed/vol6-

1/jenkins.asp

E Castronova, “Synthetic Worlds: The Business and

Culture of Online Games”, University of Chicago

Press (2005). ISBN 0-226-09626-2

N. Ellison, C. Steinfield and C. Lampe, "Spatially

Bounded Online Social Networks and Social Capital:

The Role of Facebook", Conference of the

International Communication Association (ICA), June

19-23, 2006 in Dresden, Germany

Powell, L., Richmond, V. P., & Williams, G. C. (2011).

Social networking and political campaigns:

perceptions of candidates as interpersonal constructs.

North American Journal of Psychology.

C. R.Ondrejka, “Living on the Edge: Digital Worlds

Which Embrace the Real World”, June 5, 2004.

WEBIST2012-8thInternationalConferenceonWebInformationSystemsandTechnologies

134

I. MacInnes, "The Implications of Property Rights in

Virtual World Business Models". Americas

Conference of Information Systems (AMCIS 2004),

New York, NY

G. Attwell, “IBM Centre occupied by protesting avatars”,

posted in Pontydysgu MUVEs, March 28th, 2008,

photo by RSU

D. Harry and J. Donath, "Information spaces -- building

meeting rooms in virtual environments". In CHI '08

extended abstracts on Human factors in computing

systems (CHI '08). ACM, New York, NY, USA, 3741-

3746. DOI=10.1145/1358628.1358923

P. G. Zimbardo, C. Maslach, and C. Haney, "Reflections

on the Stanford Prison Experiment: Genesis,

transformations, consquences". In T. Blass (Ed.).

Obedience to authority: Current Perspectives on the

Milgram paradigm (pp.193-237). Mahwah, N. J.:

Erlbaum, 2000

Guy, I., Zwerdling, N., Carmel, D., Ronen, I., Uziel, E.,

Yogev, S., et al. (2009). Personalized recommendation

of social software items based on social relations.

Third ACM conference on Recommender systems ,

(pp. 53-60).

J. Freyne, M. Jacovi, I. Guy and W. Geyer, “Increasing

engagement through early recommender intervention”.

Proc. RecSys’09, (2009), 85-92

D. M. Blei, A. Y Ng, M. I. Jordan and John Lafferty,

"Latent Dirichlet allocation". Journal of Machine

Learning Research 3: pp. 993-1022. January 2003

S. M. Kim, and E. H. Hovy, "Identifying and Analyzing

Judgment Opinions.". Proceedings of the Human

Language Technology / North American Association

of Computational Linguistics conference (HLT-

NAACL 2006). New York, NY

E-GOVERNMENTANDPOLICYSIMULATIONININTELLIGENTVIRTUALENVIRONMENTS

135