Towards a Cyber-physical Systems Resilience Approach based on

Artiﬁcial Emotions and Multi-agent Systems

Eskandar Kouicem, Cl

ement Ra

ıevsky and Michel Occello

Univ. Grenoble Alpes, Grenoble INP, LCIS, 26000 Valence, France

Keywords:

Artiﬁcial Emotions, Multi-agent Systems, Cyber-physical Systems, Resilience, Distributed Systems.

Abstract:

The concept of resilience is popular and studied in different domains like human and social sciences (psychol-

ogy, psychiatry, sociology etc.) and recently in cognitive science, biology, ecology and computer science. The

objective of this article is to present our research that aims to incorporate knowledge from human and social

sciences in computer science to solve cyber-physical systems resilience problems. For us humans, emotions

are considered as an important process in responding to unanticipated events, for that, emotions are important

for our resilience. Our work aims to inspire from emotional processes to create an agent model that increases

the resilience of cyber-physical systems. This agent model will integrate individual and collective processes.

In addition, one of our principal hypotheses in our research is that the multi-agent paradigm is suitable to

integrate emotion-like processes into cyber-physical systems.

1 INTRODUCTION

As humans, our emotions increase our ability to

adapt to unknown situations as individuals and as

groups (Frijda, 1986; Lazarus, 1991). So human emo-

tion is a relevant source of inspiration for solving dis-

tributed systems resilience problems that we consider

as the ability to recognize, adapt, and handle unantic-

ipated perturbations (Woods, 2012). For us, a cyber-

physical system (CPS) is a system in which its mecha-

nisms are controlled or monitored by computer-based

algorithms, composed of several subsystems and it

has at least a part of its subsystems in direct interac-

tion with the physical world. Depending on their char-

acteristics, which are the interaction with the physi-

cal environment and the open system, cyber-physical

systems can have some problems such as: incor-

rect collected data, communication losses, limited re-

sources, adding or removing components to/from the

CPS during the execution, and the unpredictability of

the encountered situations. We therefore aim to use

knowledge from psychology of emotion and sociol-

ogy to improve cyber-physical systems’ resilience, es-

pecially taking into consideration their distributed as-

pect. To this end, we adopt the multi-agent paradigm

which intrinsically and historically support using so-

cial and cognitive metaphors in the design and im-

plementation of distributed complex systems (Ferber

and Weiss, 1999). This approach will allow us to in-

tegrate cognitive science and psychology knowledge

about resilience (resilience to stress (Lazarus, 1993),

collective resilience etc.) and cognitive functions of

emotion in agents’ decision-making mechanisms and

in the self-organizing processes of the agents groups.

This paper starts by deﬁning resilience in differ-

ent domains and presents a classiﬁcation of exist-

ing resilience approaches and their problems. Af-

ter that, we present our analysis of the relevance of

emotion-related cognitive processes in our resilience

and how to draw from these cognitive processes to im-

prove resilience-related artiﬁcial processes in cyber-

physical systems. Relationships between resilience

properties and characteristics of emotion and multi-

agent systems are then proposed. Finally, we con-

clude this paper by the description of the perspective

of this work.

2 RESILIENCE

Resilience is studied by researchers in a variety of

disciplines including psychology, sociology, and re-

cently cognitive science, ecology and computer sci-

ence.

The term resilience is used differently by differ-

ent communities. Ruault et al. (2011) explain that, in

ecology, resilience is the ability of an ecosystem or

species to recover its normal behaviour after experi-

Kouicem, E., RaÃ

revsky, C. and Occello, M.

Towards a Cyber-physical Systems Resilience Approach based on Artiﬁcial Emotions and Multi-agent Systems.

DOI: 10.5220/0009176203270334

In Proceedings of the 12th International Conference on Agents and Artiﬁcial Intelligence (ICAART 2020) - Volume 1, pages 327-334

ISBN: 978-989-758-395-7; ISSN: 2184-433X

327

encing traumas. They also state that in psychology,

resilience is the ability of a person or a group to de-

velop well, to continue to project into the future, de-

spite destabilizing events, difﬁcult living conditions

and sometimes severe traumas. At the level of the

individual, traumas destroy the psyche, at the level

of the group, traumas destroy the existing bonds be-

tween the members of the group.

Generally, in systems engineering, resilience is a

rapid recovery from a degraded system state. The net-

works community deﬁnes it as the combination of re-

liability (reliability, security, performance) and toler-

ance (survivability, disturbance tolerance and trafﬁc

tolerance). The computer science community has de-

ﬁned resilience as the persistence of supplying ser-

vices and the availability of features (Trivedi et al.,

2009).

According to Woods (2012), in resilience engi-

neering, “resilience concerns the ability to recognize

and adapt to handle unanticipated perturbations that

call into question the model of competence, and de-

mand a shift of processes, strategies and coordina-

tion”.

There is often a confusion in the uses of the terms

robustness and resilience. To disambiguate the two,

we chose the deﬁnitions of (Linkov et al., 2013b):

• Robustness is the extent to which a system is able

to withstand an unexpected internal or external

event or change without degrading system perfor-

mance.

• Resilience refers to the system’s ability to re-

cover or regenerate its performance after an unex-

pected impact produces a degradation of its per-

formance (Linkov et al., 2013b).

2.1 Resilience in Cyber-physical

Systems

To position our approach relative to the cyber-

physical systems (CPSs) domain, we based our study

on the work of Lee et al. (2015) who deﬁned the 5C

architecture presented in Figure 1. This architecture

speciﬁes how a CPS is built, from the initial data ac-

quisition to the creation of ﬁnal system.

The integration of all the 5 levels in a CPS is rarely

achieved. Generally, a CPS has two main functional

components:

1. advanced connectivity that provides real-time data

acquisition from the physical world and informa-

tion about cyberspace, and

2. intelligent data management, data analysis and

computing capabilities that build cyberspace.

However, this requirement is very abstract and

not speciﬁc enough for CPS implementation in gen-

eral (Lee et al., 2015).

Since our aim is to give CPS the ability to self-

conﬁgure and self-adapt in order to improve their re-

silience, it appears that our contribution will be a part

Figure 1: 5C architecture for the implementation of a cyber-physical system (Lee et al., 2015).

ICAART 2020 - 12th International Conference on Agents and Artiﬁcial Intelligence

328

of the upmost layer of the 5C architecture: “V. Con-

ﬁguration Level”.

2.2 Existing Resilience Approaches

Existing resilience approaches are divided into two

categories: qualitative and quantitative (see Fig-

ure 2) (Hosseini et al., 2016).

• The qualitative category: it includes methods that

assess the resilience of a system without any

numerical descriptor, it also contains two sub-

categories:

– Conceptual Frameworks: offer best practices

for analyzing system’s resilience by guiding re-

silient systems principles and characteristics.

– Semi-quantitative Indices: provide expertise on

different qualitative aspects of resilience. These

approaches are developed using a set of ques-

tions designed to assess different resilience-

based characteristics (e.g., redundancy, re-

sourcefulness).

• Quantitative methods with two sub-categories:

– General Resilience Measures: determine re-

silience by comparing the system performance

before and after the disturbance without fo-

cusing on the speciﬁc system’s characteristics.

There are deterministic and stochastic mea-

sures, each one of them is used to describe the

static and dynamic system’s behaviour.

– Structural-based Approaches: examine how the

structure of a system affects its resilience by ob-

serving the system’s behaviour and modelling

or simulating the characteristics of the system.

Another classiﬁcation of existing resilience ap-

proaches into two other main categories: metric-

based and model-based approaches (see Fig-

ure 3) (Linkov and Kott, 2019).

• Metric-based Approaches: use individual com-

ponent properties or system functions measures

(metrics) to assess overall the system’s perfor-

mance. Resilience metrics are deﬁned as a mea-

surable properties of the system that quantify the

system’s objectives achievement degree.

• Model-based Approaches: use system conﬁg-

uration modeling and scenario analysis to deter-

mine overall the system’s performance. These ap-

proaches use mathematical or physical concepts

to represent the real world (environment) and de-

ﬁne the resilience. Modelling the system’s re-

silience requires the knowledge of a system’s crit-

ical functions, tasks, objectives, temporal pat-

terns, thresholds, memory, and adaptation.

Figure 3: Metric-based and model-based approaches for re-

silience assessment (Linkov and Kott, 2019).

According to Hosseini et al. (2016), the main ap-

proaches addressing the issue of resilience are either

centralized or based on redundancy. In contrast we

aim for taking the distributed nature of CPS into ac-

count at the core of the resilience-related processes

and to avoid redundancy because it is not always prac-

tically feasible in systems including physical compo-

nents.

3 PROPOSED SOLUTION

As we can see in Figure 3, agent-based approach and

multi-agent systems are considered as model-based

approaches to address resilience.

Figure 2: Classiﬁcation scheme of resilience assessment methodologies (Hosseini et al., 2016).

Towards a Cyber-physical Systems Resilience Approach based on Artiﬁcial Emotions and Multi-agent Systems

329

Our work aims to increase resilience in a cyber-

physical system by providing its subsystems with a

form of autonomy of decision that allows them to de-

tect abnormal situations and adapt their behaviour to

these situations. These mechanisms are inspired by

human emotions and allow the entire cyber-physical

system to detect abnormal situations, and then, adapt

its behaviour to these situations using artiﬁcial pro-

cesses inspired by social metaphors (Kouicem et al.,

2019). To do so, our approach utilizes the multi-agent

paradigm (Ferber and Weiss, 1999).

3.1 Resilience Proﬁle

Figure 4 presents our adaptation of the resilience pro-

ﬁle proposed by Linkov and Kott (2019). This proﬁle

represents the different resilience-related behaviours

triggered by the occurrence of an advert event.

We added a detection behaviour phase which is

active before the event and crucial to CPSs since it

is responsible for the triggering of absorption and re-

covery behaviour. Once the event is detected and the

absorption is active, the system begins the recovery

process, potentially taking into account information

about the event gathered by the detection behaviour.

Figure 4: The extended resilience proﬁle (based on (Linkov

and Kott, 2019)).

As we see, it is possible that the system will im-

prove its functionality and resilience during the adap-

tation phase. This increase is justiﬁed by the addition

of the potential gain in the detection and recovery ca-

pacity acquired during the adaptation phase to the sys-

tem’s functionality.

3.2 Multi-agent Approach for CPS

Resilience

As we have said, in our proposed solution we use a

social metaphor for cyber-physical systems. So we

have chosen the multi-agent approach because it suits

best our needs, adapted to the control of distributed

systems and addressing certain aspects of resilience.

Multi-agent systems offer us a decentralized so-

lution to solve the “single point of failure” prob-

lem inherent in centralized solutions. In this ap-

proach, each agent is autonomous and it has its own

role, there is no global control, the data and the de-

cisions are decentralized (Ferber and Weiss, 1999).

Through these characteristics a multi-agent system

can maintain its functioning in case of communica-

tion loss, decreases the volume of transmitted data,

scaling up and avoids redundancy/replication of many

software and hardware components of the cyber-

physical system using the autonomy of its agents.

Moreover, adopting the multi-agent paradigm and

self-organization (Di Marzo Serugendo et al., 2006)

breaks with traditional resilience approaches by using

a social metaphor for cyber-physical systems. Adding

to this, this paradigm allows us to use cognitive sci-

ence and psychology to reproduce the cognitive func-

tions of emotions (Ivanovi

c et al., 2015), both in

the individual’s decision-making process and in the

group’s self-organization process.

There is some work on resilience in cyber-

physical systems using a multi-agent ap-

proach (Janu

ario et al., 2018, 2019), that solves

the problem of resilience enhancement in CPSs

and WSAN (wireless sensor and actuator network)

using hierarchical multi-agent systems. They

propose frameworks based on a multi-agent ap-

proach implemented in a distributed middleware.

These frameworks aim to improve the resilience of

networked control and supervision systems. Our

proposal is largely distinct because we aim to use

artiﬁcial processes inspired by human emotions and

social metaphors to improve cyber-physical systems

resilience.

3.3 Use of Artiﬁcial Emotions for CPS

Resilience

According to Frijda (1986) emotional phenomena are:

“non-instrumental behaviours and non-instrumental

features of behaviour, physiological changes, and

evaluative subject-related experiences, as evoked by

external or mental events, and primarily by the signif-

icance of such events.”

ICAART 2020 - 12th International Conference on Agents and Artiﬁcial Intelligence

330

In computer science, artiﬁcial emotions are a set

of pre-programmed or non-scheduled processes run-

ning within a machine, facilitating decision-making

and enabling the system to adapt to the environment.

The artiﬁcial emotion is the fruit of the program’s

input-output as well as its own internal activity, and

is often the object of a collaboration with a cognitive

structure, by means of which the system deals with

the problems introduced by its environment. In ad-

dition, it is part of a programming logic more or less

explicit, but still in the ﬁeld of computable (Mahboub,

2011; Kouicem et al., 2019).

The main existing work uses artiﬁcial emotions

for human-machine interaction and the detection and

expression of natural emotions (Calvo et al., 2015).

Other work uses artiﬁcial emotions to make virtual

agents behaviours more realistic (Saunier and Jones,

2014; Adam et al., 2010). More speciﬁcally, in the

ﬁeld of cyber-physical systems, the main work con-

cerns the emotion recognition (Calero et al., 2018)

and emotional robots (Breazeal, 2003). We noticed

that this work falls within the scope of affective com-

puting (Picard, 2000).

We note that our work differs from these research

themes, it does not deal with the interaction with

humans or the simulation of natural emotions, but

rather by inspiring from their functions (Ra

ıevsky and

Michaud, 2009). In existing work on artiﬁcial emo-

tions, emotions are elicited by an analysis of sym-

bolic events related to the objectives of the system

or require a large amount of knowledge of the sys-

tem designer about the situations that the system may

encounter. These initial hypotheses of existing work

limit the resilience of the resulting systems. Further-

more, these characteristics are difﬁcult to apply to

CPSs because:

1. They are embedded in the physical world and can

not trigger emotions from symbolic data about the

situation and,

2. the distributed and opened nature of CPSs pre-

vents their designers from anticipating all possible

situations.

In order for cyber-physical systems to beneﬁt from

the capacities provided by emotions, without compro-

mising their resilience, we propose to identify generic

mechanisms, not speciﬁc to the task of the system,

to initiate the artiﬁcial equivalent of an emotional

episode. Otherwise, artiﬁcial emotions will mainly

be used to detect abnormal situations and update so-

cial organizations in response to these abnormal sit-

uations. These functions are directly related to the

“Plan/Prepare”,“Absorb” and “Adapt” phases of the

resilience proﬁle depicted in Figure 4.

The existing work that links resilience with emo-

tions is mainly in psychology and sociology (Mas-

ten et al., 1990; Lazarus, 1991, 1993; Tugade and

Fredrickson, 2004; Norris et al., 2008). We note

that our idea of creating an emotion-based approach

to increase cyber-physical systems resilience remains

original.

3.4 Resilience Matrix

In resilience proﬁle (see Figure 4), we can see the

four stages of the event management cycle that a sys-

tem needs to maintain to be resilient according to

the National Academy of Sciences (NAS) (National

Academy of Sciences, 2012; Linkov et al., 2013b,a)

which are:

1. Plan/Prepare: lay the foundation to ensure that

services and assets remain available in case of ab-

normal situations.

2. Absorb: maintain the availability of the most crit-

ical functions and services while delaying or iso-

lating the disruption.

3. Recover: restore all asset functions and service

availability to their pre-event (before the abnormal

situation) functionality.

4. Adapt: use knowledge of critical events to mod-

ify communication protocols, system conﬁgura-

tion, learning process or other aspects to increase

resilience.

Linkov et al. (Linkov et al., 2013a,b) combined those

four phases and four domains (physical, information,

cognitive and social) to create a generic matrix of

resilience metrics. In our case, we combined the

four phases of resilience with artiﬁcial emotions (cog-

nitive) and multi-agent systems (social) for obtain-

ing our resilience matrix (see Table 1). This map-

ping helped us to identify the aspects of artiﬁcial

emotions and multi-agent systems (cells of the ma-

trix) that we will integrate in our agent model to in-

crease cyber-physical systems resilience. In the ma-

trix, we used the detection phase as a combination of

“Plan/Prepare” and “Absorb” phases because it fully

covers the planning phase and trigger the absorption

phase (see Figure 4).

3.5 Scenario (An Example)

We consider a cyber-physical system composed of

several sensors and actuators as a multi-agent sys-

tem, each subsystem (sensor or actuator) of the CPS

is an agent. A sensor agent perceives data from the

environment with its sensor module. An actuator

Towards a Cyber-physical Systems Resilience Approach based on Artiﬁcial Emotions and Multi-agent Systems

331

agent acts on the environment with its actuator mod-

ule. Agents who share the same goals belong to the

same organization, whether they are sensors or actua-

tors. The agents of an organization communicate with

each other, so the actuator agents share with the other

agents their operating status and the sensor agents

provide information about the environment. So, if

an actuator agent needs some information, it asks the

sensor agents of its organization to perceive data for

it, if these ones can’t get it, they can ask the sen-

sor agents of the other organizations. Each sensor

agent stores the recently perceived data in a time se-

ries database in its own memory, then the old time

series will be shared in the memories of the actuators

of its same organization. Basically, the agents have a

predeﬁned tolerance thresholds to the perceived data

stocked in both sensor and actuator agents.

Here is a reaction scenario of the agents based on

emotional processes in case of an anomaly in the data

that a sensor agent perceives from the environment:

• Detection: the detection is always active, before

any critical event until the beginning of the ab-

sorption phase. Once a sensor agent has a doubt

about the data that it perceives by getting out-of-

domain values unusual sequences or long repeti-

tion of similar data, the planning phase begins.

• Plan/Prepare: the sensor agent increases the

sampling frequency and compares the data with a

larger time series and communicates the situation

to the other agents of his organization to conﬁrm

whether it is a true abnormal or just a false posi-

tive.

• Absorb: if the situation is conﬁrmed as abnor-

mal, the agents reduce the tolerance thresholds

and communicate this situation to other organiza-

tions of the system.

• Recover: the system’s agents continue to oper-

ate ignoring the data of this sensor agent and take

into consideration the system’s data history in the

same conditions.

• Adapt: the agents use the knowledge of this sit-

uation by modifying their communication pro-

tocol intra-organization and/or extra-organization

or updating the organizations architectures to im-

prove its resilience.

4 CONCLUSIONS AND FUTURE

WORK

In this paper, we presented our work about utilizing

the multi-agent paradigm to integrate knowledge from

psychology of emotion to improve cyber-physical

systems resilience. We presented our study of the def-

inition of resilience in relevant research domains and

presented our proposition and our choices regarding

inspiration from human emotion as well as the choice

of the multi-agent paradigm.

Our goal is to create an agent model based on

the resilience matrix that has been deﬁned. This

agent model integrates individual and decentralized

processes dedicated to the detection of abnormal sit-

uations inspired by the cognitive processes trigger-

ing emotional episodes in humans. The agent model

will also include mechanisms for adapting individual

behaviour in relation with the detected situations in

order to increase the resilience of the cyber-physical

system. Individual detection and adaptation trigger

collective processes in the system, these processes

use the information built by the individual mecha-

nisms. The individual behaviour adaptation impacts

on the organization of the agents group that controls

Table 1: The resilience matrix.

Detection phase: Plan/Prepare and

Absorb

Recover Adapt

Artiﬁcial

emotions

-Using an incremental process for

anomaly detection.

-Anticipating existing scenarios.

-Increase the sampling frequency.

-Increasing in the readiness for action.

-Decrease of tolerance thresholds.

-Rapid recovery of nominal behaviour.

-Review the response and decision-making

processes.

-Adaptation of the response behaviour of

the situation.

Multi-agent

systems

-Collective detection process.

-Propagation of detected situations.

-Cooperation for negotiate a detected

situation to decide if it is a real or a

fake abnormal situation.

-Establish a collective decision making

techniques or aids to select recovery

options.

-Sharing the responses (adaptations).

-A self-organization of the group of agents

carrying out the system’s control.

-Improve decision-making and allow

the system to adapt to the environment.

ICAART 2020 - 12th International Conference on Agents and Artiﬁcial Intelligence

332

the cyber-physical system, which will trigger self-

organization for adapting the collective behaviour to

the abnormal situation.

The proposed agent model and multi-agent sys-

tem will be tested by simulation (Jamont and Occello,

2013), and also in real cyber-physical systems in or-

der to quantify experimentally the resilience improve-

ment they bring.

ACKNOWLEDGEMENTS

This Ph.D project has received funding from the Trust

research chair of the Grenoble-INP Foundation and

the Auvergne-Rh

one-Alpes region.

REFERENCES

Adam, C., Canal, R., Gaudou, B., Vinh, H. T., Taillandier,

P., et al. (2010). Simulation of the emotion dynamics

in a group of agents in an evacuation situation. In

International Conference on Principles and Practice

of Multi-Agent Systems, pages 604–619. Springer.

Breazeal, C. (2003). Emotion and sociable humanoid

robots. International journal of human-computer

studies, 59(1-2):119–155.

Calero, J. A. M., Marino, R., Lanza-Gutierrez, J. M.,

Riesgo, T., Garcia-Valderas, M., and Lopez-Ongil, C.

(2018). Embedded emotion recognition within cyber-

physical systems using physiological signals. In 2018

Conference on Design of Circuits and Integrated Sys-

tems (DCIS), pages 1–6. IEEE.

Calvo, R. A., D’Mello, S., Gratch, J. M., and Kappas, A.

(2015). The Oxford handbook of affective computing.

Oxford University Press, USA.

Di Marzo Serugendo, G., Gleizes, M.-P., and Karageorgos,

A. (2006). Self-organisation and emergence in multi-

agent systems: An overview. Informatica, 30(1):45–

54.

Ferber, J. and Weiss, G. (1999). Multi-agent systems: an

introduction to distributed artiﬁcial intelligence, vol-

ume 1. Addison-Wesley Reading.

Frijda, N. H. (1986). The emotions. Cambridge University

Press.

Hosseini, S., Barker, K., and Ramirez-Marquez, J. E.

(2016). A review of deﬁnitions and measures of

system resilience. Reliability Engineering & System

Safety, 145:47–61.

Ivanovi

c, M., Budimac, Z., Radovanovi

c, M., Kurbalija, V.,

Dai, W., B

adic

a, C., Colhon, M., Ninkovi

c, S., and

Mitrovi

c, D. (2015). Emotional agents–state of the art

and applications. Computer Science and Information

Systems, 12(4):1121–1148.

Jamont, J.-P. and Occello, M. (2013). Using mash in the

context of the design of embedded multiagent system.

In International Conference on Practical Applications

of Agents and Multi-Agent Systems, pages 283–286.

Springer.

Janu

ario, F., Cardoso, A., and Gil, P. (2018). Multi-agent

framework for resilience enhancement over a wsan.

In 2018 15th International Conference on Electrical

Engineering/Electronics, Computer, Telecommunica-

tions and Information Technology (ECTI-CON), pages

110–113. IEEE.

Janu

ario, F., Cardoso, A., and Gil, P. (2019). A distributed

multi-agent framework for resilience enhancement

in cyber-physical systems. IEEE Access, 7:31342–

31357.

Kouicem, E., Ra

ıevsky, C., and Occello, M. (2019). Artiﬁ-

cial Emotions for Distributed Cyber-physical Systems

Resilience. In Proceedings of the Cyber-Physical Sys-

tems PhD Workshop 2019, pages 84–95.

Lazarus, R. S. (1991). Emotion and adaptation. Oxford

University Press on Demand.

Lazarus, R. S. (1993). From psychological stress to the

emotions: A history of changing outlooks. Annual

review of psychology, 44(1):1–22.

Lee, J., Bagheri, B., and Kao, H.-A. (2015). A cyber-

physical systems architecture for industry 4.0-based

manufacturing systems. Manufacturing Letters, 3:18–

23.

Linkov, I., Eisenberg, D. A., Bates, M. E., Chang, D., Con-

vertino, M., Allen, J. H., Flynn, S. E., and Seager, T. P.

(2013a). Measurable resilience for actionable policy.

Environmental Science & Technology, 47(18):10108–

10110.

Linkov, I., Eisenberg, D. A., Plourde, K., Seager, T. P.,

Allen, J., and Kott, A. (2013b). Resilience metrics for

cyber systems. Environment Systems and Decisions,

33(4):471–476.

Linkov, I. and Kott, A. (2019). Fundamental concepts of

cyber resilience: Introduction and overview. In Cy-

ber resilience of systems and networks, pages 1–25.

Springer.

Mahboub, K. (2011). Mod

elisation des processus

emotionnel dans la prise de d

ecision.(Emotional pro-

cesses modelling in decision making). PhD thesis,

University of Le Havre, France.

Masten, A. S., Best, K. M., and Garmezy, N. (1990).

Resilience and development: Contributions from the

study of children who overcome adversity. Develop-

ment and psychopathology, 2(4):425–444.

National Academy of Sciences, N. (2012). Disaster re-

silience: A national imperative. Washington, DC: The

National Academies Press.

Norris, F. H., Stevens, S. P., Pfefferbaum, B., Wyche, K. F.,

and Pfefferbaum, R. L. (2008). Community resilience

as a metaphor, theory, set of capacities, and strategy

for disaster readiness. American journal of community

psychology, 41(1-2):127–150.

Picard, R. W. (2000). Affective computing. MIT press.

ıevsky, C. and Michaud, F. (2009). Emotion generation

based on a mismatch theory of emotions for situated

agents. In Handbook of Research on Synthetic Emo-

tions and Sociable Robotics: New Applications in Af-

Towards a Cyber-physical Systems Resilience Approach based on Artiﬁcial Emotions and Multi-agent Systems

333

fective Computing and Artiﬁcial Intelligence, pages

247–266. IGI Global.

Ruault, J.-R., Luzeaux, D., Colas, C., and Sarron, J.-

C. (2011). R

esilience des syst

emes sociotechniques

application

a l’ing

enierie syst

eme. G

enie logiciel,

96:40–52.

Saunier, J. and Jones, H. (2014). Mixed agent/social dy-

namics for emotion computation. In Proceedings

of the 2014 international conference on Autonomous

agents and multi-agent systems, pages 645–652. In-

ternational Foundation for Autonomous Agents and

Multiagent Systems.

Trivedi, K. S., Kim, D. S., and Ghosh, R. (2009). Resilience

in computer systems and networks. In Proceedings

of the 2009 International Conference on Computer-

Aided Design, pages 74–77. ACM.

Tugade, M. M. and Fredrickson, B. L. (2004). Resilient in-

dividuals use positive emotions to bounce back from

negative emotional experiences. Journal of personal-

ity and social psychology, 86(2):320.

Woods, D. D. (2012). Essential characteristics of resilience.

In Resilience engineering, pages 33–46. CRC Press.

ICAART 2020 - 12th International Conference on Agents and Artiﬁcial Intelligence

334