A Reﬂective Architecture for Agent-Based Models Applied to Social

Network Sites

Diego Nu

nez, Tom

as V

elez, Paul Leger

and Daniel San Mart

ın

School of Engineering, Catholic University of the North, Larrondo 1281, Coquimbo, Chile

Keywords:

Word-of-Mouth, Agent-Based Model, Simulation, Social Media, Open Implementations, Tower Reﬂection.

Abstract:

Social network sites serve as effective platforms for word-of-mouth marketing (WOM), often analyzed through

Agent-Based Models (ABMs). However, implementing ABMs can be daunting, with programmers facing

the choice of building from scratch or using frameworks. To tackle this, we propose FASOW (Flexible

Agent Simulator for Open WOM) architecture, employing the Reﬂective Tower design. FASOW’s four

layers cater to varying complexities, simplifying implementation by breaking down models into manageable

sub-layers. We validate FASOW through a case study on Twitter, examining agent saturation effects in

WOM marketing. Results indicate FASOW’s efﬁcacy, though further use cases are needed for comprehensive

evaluation. Additionally, we offer an online proof-of-concept for this architecture.

1 INTRODUCTION

People seek opinions from relatives, friends,

or experts about products before purchasing,

both ofﬂine and online (Kaplan and Haenlein,

2010). This form of communication, known as

Word-Of-Mouth (WOM), is highly valued for its

credibility and inﬂuence (Hennig-Thurau et al.,

2004). Consequently, companies aim to leverage

WOM to organically shape public perception of

their products or services (Kozinets et al., 2010).

WOM offers ease and convenience in information

sharing for both customers and businesses (Gupta

and Harris, 2010; Hennig-Thurau et al., 2015; Jansen

et al., 2009). Recognizing its signiﬁcance, WOM has

emerged as a potent marketing tool, referred to as

word-of-mouth marketing (Gupta and Harris, 2010).

This strategy involves disseminating a message about

a product or service to a group of initial customers

(referred to as ”seeds”), with the aim of encouraging

them to further propagate the message (Shen and

Hahn, 2007).

An Agent-Based Model (ABM) (Goldenberg

et al., 2001; Rand and Rust, 2011; Grimm et al.,

2006) is a technique that allows researchers to model

real-world situations by abstracting and identifying

only the crucial elements to study a phenomenon. By

https://orcid.org/0000-0003-0969-5139

https://orcid.org/0000-0001-5274-0148

using this technique, we can simulate and analyze

the behavior and the interactions that take place

among agents in a determined environment. Agents

can be considered as individuals who follow a

behavior, have a deﬁned and mutable state, and can

interact with other agents. By using ABMs, we

can model a real-world-like representation of social

network sites, where agents can share messages that

can be read by other agents and where agents can

inﬂuence decisions of other agents. Thanks to ABMs,

researchers in WOM marketing can test hypothetical

scenarios achieving more success when implementing

marketing strategies (L

opez et al., 2023; Araya et al.,

2019).

Two approaches exist for implementing

agent-based simulations: building from scratch

or using frameworks like NetLogo (Wilensky, 2023),

Repast (North et al., 2013), or MASON (Luke

et al., 2005). Building from scratch is complex

due to the need to handle numerous abstractions,

while frameworks offer pre-built solutions but may

require learning and customization, especially with

Repast. Regardless of the approach chosen, deep

knowledge of software development is necessary. To

address these challenges, the FASOW architecture

is proposed, leveraging the Reﬂective Tower design

strategy. This architecture organizes layers to

facilitate the gradual implementation of complex

ABM models, providing ﬂexibility as complexity

increases.

Nuñez, D., Vélez, T., Leger, P. and San Martín, D.

A Reﬂective Architecture for Agent-Based Models Applied to Social Network Sites.

DOI: 10.5220/0012615100003690

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 26th International Conference on Enterprise Information Systems (ICEIS 2024) - Volume 1, pages 981-988

ISBN: 978-989-758-692-7; ISSN: 2184-4992

981

We validated the FASOW architecture through

one case that applied to the Twitter site. This

case is presented in (L

opez et al., 2023) to analyze

the saturation effect on agents when repeatedly

interacting with a WOM advertising message. Our

paper aims to develop a software architecture

for Agent-Based Models (ABMs), making use of

abstractions present in programming languages to

facilitate the learning and use of the software during

the implementation of ABM models on WOM

marketing campaigns.

The paper is organized as follows. Section 2

presents the concepts that are required to understand

FASOW. Section 3 describes the FASOW architecture

for the simulation of WOM marketing campaigns.

Section 4 discusses two frameworks to simulate

ABM (Repast Simphony, Simudyne) with FASOW.

Section 5 introduces a preliminary validation aimed

at testing the capabilities of our architecture. This

validation is conducted on a pre-existing model

focused on WOM marketing campaigns for Twitter.

Section 6 presents the conclusions and its current

limitations.

2 BACKGROUND

This section establishes the fundamental concepts

necessary for a clear comprehension of our proposal.

It is divided into three subsections: An Introduction

explaining what WOM marketing campaigns are and

how they work, a Description of agent-based models

and their applications to WOM marketing campaigns

and the reﬂection tower concept.

2.1 WOM-Based Marketing Campaigns

When a consumer is going to decide to purchase a

new product, they must ﬁrst be aware of its existence

and its characteristics (Iyengar et al., 2011). Several

marketing studies indicate that to create awareness

about the existence of a product, it is essential to

spread information about it (L

opez and Sicilia, 2013),

which is why word-of-mouth (WOM) has become a

very useful tool for companies (Gupta and Harris,

2010). We can ﬁnd an electronic version of this

type of communication through the Internet, which is

called (eWOM) electronic word-of-mouth (L

opez and

Sicilia, 2014). This type of communication is deﬁned

as any opinion spread over the Internet by consumers

about any product or company (Hennig-Thurau et al.,

2004).

The rise of electronic Word-of-Mouth (eWOM)

has been fueled by the vast amount of information

shared online, empowering consumers to exchange

insights via social networking platforms (Cheung

and Thadani, 2012). This digital discourse

allows individuals to form opinions and make

purchasing decisions without traditional face-to-face

interactions. Social environments, including friends

and acquaintances, directly inﬂuence consumer

choices, alongside opinions shared by fellow

consumers online. For instance, product reviews

from experienced users can signiﬁcantly sway

decisions.

When companies encourage consumers to

spread information about their products, it’s

termed Word-of-Mouth (WOM) marketing (Araya

et al., 2019). This approach has become a vital

communication channel for companies, often

initiated through strategic campaigns involving

diverse “seed” consumers. These seeds, possessing

varied characteristics, are pivotal in effectively

disseminating targeted information. Research

has identiﬁed different user types, such as “hub”

users with extensive connections, who accelerate

information spread and impact market size (Libai

et al., 2013). Studies have explored various WOM

campaign strategies, including the inﬂuence of

distinct consumer types on content dissemination

through platforms like retweet (da Chen et al., 2002;

Hinz et al., 2011).

2.2 Agent-Based Model in WOM

Marketing Campaigns

An Agent Based-Model (ABM) (Leger et al.,

2016) simulates the behavior of the members of

a population, which change due to the interaction

that each of them has with its neighbors (Neumann

and Burks, 1966). It is a model in which, in an

environment, there are agents or individuals with a

ﬁnite number of possible states which are updated

based on discrete time intervals. A simulation has

advantages over similar methods that also study the

relationship between entities (Araya et al., 2019),

such as: (1) describing complex systems from simple

real-world rules; (2) the possibility of simulating

interactions between entities, regardless of the lack

or limitation of data; (3) the ease of testing different

hypothetical scenarios by modifying the variables

present in the computational environment of the

simulation; and (4) a low-cost methodology, since its

validity does not lie in the amount of data but in the

strength of the theory. The main components of an

agent-based simulation model are:

Environment. Corresponds to the space where agents

interact with others (Goldenberg et al., 2001). This

ICEIS 2024 - 26th International Conference on Enterprise Information Systems

982

environment can simulate different scenarios, either

online, such as a social network, or ofﬂine (ofﬂine),

such as a network of neighbors.

Agent. Agents represent individuals or some entity

with a state and behavior capable of interacting with

its environment. They have different properties and

behaviors (Rand and Rust, 2011).

Action. Agents execute actions that will eventually

affect others, for each time interval based on the

previously deﬁned set of rules (Libai et al., 2013). For

example, in a social network, an agent may take the

action of sharing certain information.

Rules. Given an agent in a speciﬁc environment

and time, the rules deﬁne the dynamics of how the

model (Araya et al., 2019) will evolve. These can be

probabilistic or deterministic and can vary between

agents.

Period. For each period, an agent can perform one

action, and each period represents a speciﬁc discrete

time (Libai et al., 2013).

Using ABMs to model WOM marketing

campaigns (Chica and Rand, 2017) is useful since

it does not require a large amount of data, but only

the necessary to calibrate and validate the model.

Therefore, it is not a requirement for researchers to

know the emergent behaviors of a population, but

only individual knowledge of each agent is necessary.

An ABM is appropriate when an environment is

complex and dynamic, just like a social network.

Mainly when the marketing interest is a result of

interactions between consumers (Chica and Rand,

2017).

2.3 Reﬂection Tower

A reﬂection tower is a reﬂective system that is formed

by N levels of segmentation. Each level contains its

degree of introspection on a speciﬁc program. This

concept can be decomposed into two processes, called

reiﬁcation and reﬂection, both of which correspond to

the process of moving up or down the tower (Ibrahim,

1992).

Reiﬁcation is the process where the current

working level is manipulated and transformed so that

a higher level can work with it. Reﬂection is the

process by which the data corresponding to the higher

level is reinstated at a lower level. Introspection is

the ability of a program to reason about its reiﬁcation

or another aspect of itself. Therefore, the reﬂective

structure denotes a program’s capacity to access its

representation and structure, while reﬂective behavior

refers to a program’s dynamic ability to access

itself (Ibrahim, 1992). Keeping these concepts in

Instrospection Level N

Instrospection Level 3

Instrospection Level 2

Instrospection Level 1

(Base Interface)

Meta Interface

Developer

Reification

Reflection

Figure 1: Reﬂection Tower and Open Implementations.

mind, reﬂection can be viewed as the capability

to inspect and/or manipulate the implementation

structures of other systems utilized by a program.

Figure 1 schematically depicts the reﬂection tower

and illustrates the concept of an open implementation.

As deﬁned in (Kiczales et al., 1997), an open

implementations is characterized by a system offering

a minimum of two interconnected interfaces to

its clients: 1) a base-level interface providing

access to the system’s core functionality and 2) a

meta-level interface that discloses certain aspects of

the implementation underlying the base interface.

In these principles, the reﬂection tower serves as

a mechanism to achieve an open implementation,

enabling the exposition of the implementation details

of speciﬁc components to users.

A reﬂective implementation of Agent-Based

Models (ABMs) can encompass a multitude of

interconnected concepts. For example, adopting the

reﬂection tower offers a systematic methodology

for understanding and incrementally implementing

Agent-Based Modeling (ABM) concepts. This

involves categorizing the concepts related to an

ABM model into distinct levels. The reﬂective

implementation becomes achievable through this

segmented, level-based approach, where each

introspection level supports a speciﬁc module linked

to the ABM.

3 FASOW: A REFLECTIVE

ARCHITECTURE FOR ABMs IN

SOCIAL NETWORKS

This section introduces our approach, named FASOW

(Flexible Agent Simulator for Open WOM), a

reﬂective architecture for the simulation of social

networking sites for WOM marketing based on

A Reﬂective Architecture for Agent-Based Models Applied to Social Network Sites

983

the reﬂection tower of the open implementation in

programming languages (Demers and Malenfant,

1995).

The development of the FASOW architecture

follows the guidelines for conceptually describing

Agent-Based Models (ABMs), as articulated

by (Grimm et al., 2006). In addition to adhering to

these recommendations, the architecture integrates

the concept of the ”reﬂection tower,” a pervasive

motif found in programming languages such as

3-Lisp, Java, or Smalltalk (Malenfant et al., 1996).

Furthermore, the implementation capitalizes on

well-established software design patterns, with the

Factory pattern (Gamma et al., 1993) serving as an

exemplar.

In FASOW, the reﬂection tower was designed to

implement an open architecture to support different

models of WOM marketing ABMs, which formalizes

the design of these ABMs and segments the models

into complexity levels. As the complexity of the

features to be added increases, the designer must

move through the different levels of the reﬂection

tower to add speciﬁc modules that represent the

changes he needs. By adopting a design approach

that segments complexity based on the features to

be added, the architecture encourages a structured

hierarchy.

Each level of this reﬂection tower serves as an

accessible entry point for developers. This tiered

design not only facilitates a systematic understanding

of the software but also offers scalability by enabling

the gradual addition of modules. As the need arises

to implement more speciﬁc features in a model, the

reﬂection tower becomes a powerful tool, allowing

developers to access varying levels of complexity to

obtain speciﬁc perspectives on the problem. This

streamlined approach simpliﬁes both comprehension

and implementation of the model.

As follows, we present the FASOW architecture

and a description of FASOW using the ODD Protocol.

3.1 FASOW Architecture

Figure 2 introduces the FASOW architecture.

The layered, tower-like structure of FASOW

facilitates diverse levels of abstraction, enabling

the simulation of WOM marketing models tailored

for social networks. The FASOW layers are (1)

Experiment, (2) Environment, (3) Agent, (4)

Action. Additionally, we can ﬁnd three modules:

a meta-programming module that controls the

reﬂection tower called TowerHandler, a module to

generate and control the output of the simulation

called DataHandler, and a module that handles

the time and the repetitions of the simulation,

that is called TimeKeeper. In the following, the

functionality and objective of the tower layers are

explained.

Action

Agent

Environment

Experiment

TimeKeeper

FASOW

DataHandler TowerHandler

create()

select()

QueryData/SelectOutput

QueryData/

SelectOutput

Figure 2: FASOW Architecture.

Experiment. This layer enables the simulation of a

model using a familiar language to people immersed

in WOM marketing.

When introducing new features that need

modiﬁcations to the core FASOW structure,

adjustments should be initially made in the

Experiment layer. Subsequently, as required,

modiﬁcations should be extended to the upper layers

of the reﬂection tower using the TowerHandler.

Environment. This layer allows users to set up the

behavior of the simulation during the step-by-step

running.

Agent. This layer allows users to create, group,

and combine different types of agents. It also

allows users to relate the agents with their respective

conﬁgurations and their set of possible actions, as

well as to establish the order of execution of these.

Action. This layer allows the creation of new actions.

Actions are functions that are executed by agents, and

handle how they react when receiving a message.

Data Handler. The DataHandler module allows

developers to deﬁne the output to be expected at the

end of the simulation.

Tower Handler. TowerHandler module is an API for

interacting between tower layers. It uses the Facade

pattern to hide the complexity behind the levels of

abstraction, acquiring the ability to intervene in the

ICEIS 2024 - 26th International Conference on Enterprise Information Systems

984

Figure 3: Our proposal, FASOW, merges the guidelines for describing ABMs proposed by Grimm et al. (Grimm et al., 2006)

and developing adaptive software introduced by Kiczales et al. (Kiczales et al., 1997).

deep levels of the tower and communicate them, this

without having to know everything behind each level.

TimeKeeper. This module handles time in the

simulation, allowing to set the maximum time to run

the simulation and the maximum number of times that

can be repeated.

In the FASOW architecture, the TimeKeeper

informs the DataHandler when a simulation step

concludes, facilitating the collection of data to

generate the ﬁnal output.

3.2 FASOW Description

This subsection explains the description of FASOW

through the Overview Design concepts and Details

(ODD) protocol outlined in (Grimm et al., 2006).

The ODD protocol delineates three distinct groups

of elements for constructing an Agent-Based Model

(ABM): (1) Overview, (2) Design concepts, and

(3) Details, as illustrated in Figure 3. The

ﬁrst element articulates the purpose and offers a

concise introduction to the model. The second

element deﬁnes the operational mechanics essential

for the ABM under construction. The ﬁnal element

delves into the speciﬁcs of the ABM, encompassing

details such as initialization and input parameters.

Subsequent sections will delve into a comprehensive

exploration of these three integral elements.

Overview. As we see in the previous subsection 3.1,

the Experiments in FASOW correspond to an

implementation of a model of a WOM marketing

campaign model in a Social Network Site (SNS)

where the agents of the model are the users of the

site. These agents follow rules that are deﬁned by the

marketing campaign of the model that is simulated.

For all cases, FASOW uses a discrete quantity of time,

which increases one by one, to handle the time in the

simulation. The description of the architecture of this

element can see graphically in Figure 3.

Design Concepts. We operate within the realm of

WOM marketing, with SNSs detailed in subsection

2.1. To encapsulate this context within our

architecture, we previously deﬁned the Environment

as the representation of the SNS where the WOM

campaign unfolds. This Environment serves as a

foundation for managing Agents, which, in turn,

embody the users involved in the WOM campaign.

The interaction between the Environment

and the Agents delineates the dynamics of user

communication during information sharing. For

agents to participate in a WOM campaign, they

must adhere to speciﬁc rules, the occurrence of

which is contingent upon predeﬁned conditions. This

introduces stochastic in the campaign results, as

the occurrence of these rules is probabilistic. The

impact on the model is managed by repeating the

Experiment, as detailed in Subsection 3.1. A visual

representation of the concept described above is

presented in Figure 3.

Details. The initialization of the Experiment is

performed by the loading of the conﬁguration that

was deﬁned in the strategy. After that loading, the

TowerHandler starts to create the Environment as a

Social Network Site (SNS), the Agents as the users of

the network, and their Actions as a part of the rules

of the campaign. Done that, the followers are added,

ending with the initialization of the TimeKeeper in

zero. The DataHandler can manage part of the

dynamic variables that can vary in the simulation, like

the state of some agent, by decorating that variable as

part of the output. Finally, it is essential to clarify

that FASOW does not support the seeds conﬁguration

to deﬁne the randomness. These modules are part of

this element, as seen in Figure 3.

4 RELATED WORK

This section compares FASOW to other

ABM frameworks: Repast Symphony (North

A Reﬂective Architecture for Agent-Based Models Applied to Social Network Sites

985

Figure 4: WOM Communication Process.

et al., 2013), Simudyne (Simudyne, 2023) and

SocLab (Rodr

ıguez Zoya et al., 2018). These

frameworks will be described and contrasted against

FASOW.

Repast Symphony. This software is an open software

that is used for implementing ABMs, and it is highly

conﬁgurable. In Repast, agents belong to a context,

and the context contains projections, where agents

live. Their projections can be graphs, matrices,

or even three-dimensional spaces. This allows for

managing an agent’s physical location at a certain

point in a simulation.

Simudyne. This software is also used for

implementing ABMs, especially ones based on

graphs, since it allows the creation of connections

and relationships between agents, which in this case

are nodes, creating a topology. The communication

between agents is based on “messages”; therefore,

it uses message-oriented programming. One of the

critical features of Simudyne is the use of decorators

for the output. This means that the information can

be processed or changed for each iteration, enabling

the possibility of more clean and comprehensive

production.

SocLab. is a software that allows modeling different

types of social organizations and studying the

interaction between agents. SocLab’s purpose is

to formalize social theories in a meta-model for

modeling and social simulation of organizations.

It is based on entities called Organizations,

containing Actors and Relationships, which interact

under Controls, Stakes, Effects, Solidarities, and

Constraints.

FASOW. Our software primarily focuses on

simulating WOM marketing campaigns through

ABMs, while the alternatives provide a more

extended range of purposes. While this might seem

a disadvantage, FASOW allows developers to run

experiments based on already-deﬁned models for

simulating social sites, thus removing the necessity

of creating extra classes that implement social site

behaviors. Furthermore, since the software is very

ﬂexible, developers can gradually implement changes

on each level of the Reﬂective Tower, simplifying the

implementation according to what is needed. Like

Simudyne, FASOW also uses decorators to produce

the experiment’s output.

A concrete implementation of FASOW

is available on https://github.com/sasow-org/

fasow-monorepo (rev. 4283028), and an

online proof-of-concept of our proposal is on

https://sasow-org.github.io/fasow-monorepo.

5 VALIDATION

Recalling our approach, Figure 4 illustrates the four

stages comprising the FASOW simulation process

as part of the validation proposal. The initial

stage is dedicated to overseeing repetitions and

ICEIS 2024 - 26th International Conference on Enterprise Information Systems

986

initializing the Experiment. This involves loading

the conﬁguration speciﬁed within it and creating

the necessary agents. Subsequently, the following

stage initiates the simulation, dispatching WOM

marketing campaign messages from the seed agents,

and executing the iterations.

The penultimate stage takes charge of

handling message reception and their subsequent

retransmission through the execution of agent

actions. Ultimately, in the concluding stage, the

iteration wraps up, and the system processes the

culmination of this cycle. It proceeds to generate

the row corresponding to the output for this

period, setting the stage for the repetition of the

entire process. In order to validate our proposal,

we implemented an ABM that was previously

published (L

opez et al., 2023). This implementation

shows how the use the capabilities of our architecture

with its reﬂective tower. The ABM is about how

the repetition of the messages in a WOM marketing

campaign can saturate agents. The impact of message

repetition (L

opez et al., 2023) examines saturation

effects on agents in WOM marketing campaigns

where messages are repeatedly circulated. Seed

agents initiate and potentially resend messages

multiple times, inﬂuencing other agents’ perceptions

of the advertised product or service.

Figure 5 illustrates architecture modiﬁcations for

implementing this model, spanning all layers.

Figure 5: Case: Message Repetition Implementation.

This entails integrating specialized modules and

extending FASOW’s Twitter implementation. New

components include an Experiment for simulation, an

Environment supporting message repetition, updated

Agents managing saturation and repetition, and

Actions triggering agent saturation upon repeated

message exposure. Simulation setup involves

registering modules with the TowerHandler and

conﬁguring Experiment parameters like action

probabilities, agent population percentage, message

repetitions, and maximum TimeKeeper ticks. This

exempliﬁes the modular thinking facilitated by the

reﬂective tower. It prompts developers to consider

Experiment creation ﬁrst, followed by Environment

setup, Agent behavior, and Actions management,

aligning with the reﬂective tower’s structured

approach.

6 CONCLUSION

Word-of-mouth (WOM) marketing is a way to share

information between people through a conversation

about some product or service, where the consumer

opinion about the product can be inﬂuenced. An

electronic version of this marketing (eWOM) can

be found on social network sites (SNSs), where

people share their opinions on these sites. An

agent-based Model (ABM) is a technique that

allows the creation and study of models that

represent a real-world phenomenon. Therefore,

WOM marketing researchers can use ABM to test

hypothetical scenarios and could be more assertive

when implementing a marketing strategy in the real

world.

We propose a Flexible Agent Simulator for Open

WOM (FASOW), a software architecture that allows

the implementation of ABMs of WOM marketing in

SNSs. This architecture is based on the strategy of

the concept of reﬂective tower, which was presented

in programming languages. Our proposal allows

developers to open implementation details to adjust

the speciﬁc and potentially unforeseen needs of ABM

models for SNSs. Additionally, FASOW facilitates a

gradual implementation of complex features through

layers. These layers represent the concept of a

reﬂection tower, which helps to divide an ABM model

into modules that are being progressively added. This

gradualness in how the modules are added is directly

proportional to the complexity of the model.

REFERENCES

Araya, N., Leger, P., and Lopez, M. (2019). Whom

do i Choose to Diffuse Information onTwitter? An

Agent-Based Model Approach. IEEE Latin America

Transactions, 17(4):677–683.

Cheung, C. M. and Thadani, D. R. (2012). The

impact of electronic word-of-mouth communication:

A literature analysis and integrative model. Decision

Support Systems, 54(1):461–470.

Chica, M. and Rand, W. (2017). Building Agent-Based

Decision Support Systems for Word-of-Mouth

A Reﬂective Architecture for Agent-Based Models Applied to Social Network Sites

987

Programs: A Freemium Application. Journal of

Marketing Research, 54(5).

da Chen, L., Gillenson, M. L., and Sherrell, D. L.

(2002). Enticing online consumers: An extended

technology acceptance perspective. Information and

Management, 39(8):705–719.

Demers, F.-N. and Malenfant, J. (1995). Reﬂection in

logic, functional and object-oriented programming: a

short comparative study. In Proceedings of the IJCAI,

volume 95, pages 29–38.

Gamma, E., Helm, R., Johnson, R., and Vlissides, J.

(1993). Design patterns: Abstraction and reuse

of object-oriented design. In European Conference

on Object-Oriented Programming, pages 406–431.

Springer.

Goldenberg, J., Libai, B., and Muller, E. (2001). Talk

of the Network: A Complex Systems Look at the

Underlying Process of Word-of-Mouth. Marketing

Letters, 12(3):211–223.

Grimm, V., Berger, U., Bastiansen, F., Eliassen, S., Ginot,

V., Giske, J., Goss-Custard, J., Grand, T., Heinz,

S. K., Huse, G., Huth, A., Jepsen, J. U., Jørgensen,

C., Mooij, W. M., M

uller, B., Pe’er, G., Piou, C.,

Railsback, S. F., Robbins, A. M., Robbins, M. M.,

Rossmanith, E., R

uger, N., Strand, E., Souissi, S.,

Stillman, R. A., Vabø, R., Visser, U., and DeAngelis,

D. L. (2006). A standard protocol for describing

individual-based and agent-based models. Ecological

Modelling, 198(1):115–126.

Gupta, P. and Harris, J. (2010). How e-wom

recommendations inﬂuence product consideration

and quality of choice: A motivation to process

information perspective. Journal of Business

Research, 63(9):1041–1049.

Hennig-Thurau, T., Gwinner, K. P., Walsh, G., and

Gremler, D. D. (2004). Electronic word-of-mouth

via consumer-opinion platforms: What motivates

consumers to articulate themselves on the Internet?

Journal of Interactive Marketing, 18(1):38–52.

Hennig-Thurau, T., Wiertz, C., and Feldhaus, F. (2015).

Does twitter matter? the impact of microblogging

word of mouth on consumers’ adoption of new

movies. Journal of the Academy of Marketing Science,

43:375–394.

Hinz, O., Skiera, B., Barrot, C., and Becker, J. U. (2011).

Seeding strategies for viral marketing: An empirical

comparison. Journal of Marketing, 75(6):55–71.

Ibrahim, M. H. (1992). Reﬂection in object-oriented

programming. International Journal on Artiﬁcial

Intelligence Tools, 1(01):117–136.

Iyengar, R., van den Bulte, C., and Valente, T. W. (2011).

Opinion leadership and social contagion in new

product diffusion. Marketing Science, 30(2):195–212.

Jansen, B. J., Zhang, M., Sobel, K., and Chowdury,

A. (2009). Twitter power: Tweets as electronic

word of mouth. Journal of the American

Society for Information Science and Technology,

60(11):2169–2188.

Kaplan, A. M. and Haenlein, M. (2010). Users of the world,

unite! The challenges and opportunities of Social

Media. Business Horizons, 53(1):59–68.

Kiczales, G., Lamping, J., Lopes, C. V., Maeda, C.,

Mendhekar, A., and Murphy, G. (1997). Open

implementation design guidelines. In Proceedings

of the 19th international conference on Software

engineering, pages 481–490, Boston, United States.

Kozinets, R. V., Valck, K. D., Wojnicki, A. C., and Wilner,

S. J. (2010). Networked narratives: Understanding

word-of-mouth marketing in online communities.

Journal of Marketing, 74(2):71–89.

Leger, P., L

opez, M., Hildago-Alc

azar, C., and Fukuda, H.

(2016). An open agent-based model to simulate the

effect of wom marketing campaigns. EAI Endorsed

Transactions on Smart Cities, 1(4).

Libai, B., Muller, E., and Peres, R. (2013). Decomposing

the value of word-of-mouth seeding programs:

Acceleration versus expansion. Journal of Marketing

Research, 50(2):161–176.

opez, M., Hidalgo-Alc

azar, C., and Leger, P. (2023).

The effect of message repetition on information

diffusion on twitter: An agent-based approach. IEEE

Transactions on Professional Communication, pages

150–169.

opez, M. and Sicilia, M. (2013). How WOM marketing

contributes to new product adoption: Testing

competitive communication strategies. European

Journal of Marketing, 47(7):1089–1114.

Luke, S., Ciofﬁ-Revilla, C., Panait, L., Sullivan, K., and

Balan, G. (2005). Mason: A multiagent simulation

environment. SIMULATION, 81(7):517–527.

opez, M. and Sicilia, M. (2014). Determinants of

E-WOM inﬂuence: The role of consumers’ internet

experience. Journal of Theoretical and Applied

Electronic Commerce Research, 9(1):28–43.

Malenfant, J., Jacques, M., and Demers, F. N. (1996).

A tutorial on behavioral reﬂection and its

implementation. In Proceedings of the Reﬂection,

volume 96, pages 1–20.

Neumann, J. V. and Burks, A. W. (1966). Theory of

Self-Reproducing Automata. University of Illinois

Press, USA.

North, M. J., Collier, N. T., Ozik, J., Tatara, E. R., Macal,

C. M., Bragen, M., and Sydelko, P. (2013). Complex

adaptive systems modeling with repast simphony.

Complex Adaptive Systems Modeling, 1(1).

Rand, W. and Rust, R. T. (2011). Agent-based modeling in

marketing: Guidelines for rigor. International Journal

of Research in Marketing, 28(3):181–193.

Rodr

ıguez Zoya, L. G., Aguirre, J. L., and Roggero, P.

(2018). Using SocLab for modeling and simulating

political clientelism.

Shen, W. and Hahn, J. (2007). Impact of online

word-of-mouth on the market for consumer goods:

The interplay between adoption rate, product market

life and market size. In ACR Pre-Conference on

Consumer Behavior Online.

Simudyne (2023). Enterprise Simulation Software: AI

Technology (v.2.6). https://simudyne.com.

Wilensky, U. (2023). Netlogo: A multi-agent

programmable modeling environment (v.2.11).

https://ccl.northwestern.edu/netlogo.

ICEIS 2024 - 26th International Conference on Enterprise Information Systems

988