TOWARDS USING DEVS FOR MODELLING ADAPTIVE STORYTELLING IN

VIRTUAL GAMES

Aznam Yacoub Gabriela Nicolescu

Aix Marseille Université, Ecole Polytechnique de Montreal,

Université de Toulon, CNRS, LIS Université de Montreal
MoFED HESL
Marseille, France Montreal, QC, Canada
gabriela.nicolescu@polymtl.ca
Ecole Polytechnique de Montreal
HESL
Montreal, QC, Canada
aznam.yacoub@polymtl.ca

Maamar el-Amine Hamri Claudia Frydman

Aix Marseille Université, Aix Marseille Université,

Université de Toulon, CNRS, LIS Université de Toulon, CNRS, LIS
MoFED MoFED
Marseille, France Marseille, France
amine.hamri@lis-lab.fr claudia.frydman@lis-lab.fr

ABSTRACT
For the last decades, Verification and Validation techniques have been well improved in order to make
safer complex systems. Sophisticated algorithms and methodologies have been proposed in the domain
of formal modelling, and simulation. However, like for testing approaches, all these methodologies suffer
from a strong weakness, while they depend on specifications, use cases and test scenarios. Indeed, if the
experimental frame is not well-defined, modelling and simulation necessarily lack in accuracy. To overcome
this challenge, a solution consists of automatically generate validated scenarios during the simulation, by
considering the experimental frame also as a simulation model. Because this problem is similar to the
adaptation of the stories in virtual worlds, this paper propose to explore a way to model adaptive storytelling
in virtual games using the DEVS formalism.
Keywords: adaptive scenarios, virtual games, story generation, DEVS, interactive storytelling.

1 INTRODUCTION
Nowadays, systems integrate more and more different kinds of complex multimodal interactions, between
the components which compose them, and also with the actors which interacts with them. The complexity

SpringSim-SCSC, 2018 July 9-12, Bordeaux, France; ⃝2018

c Society for Modeling & Simulation International (SCS)

358
 Yacoub, Nicolescu, Hamri, and Frydman

of these interactions grows with the size of the global systems, which makes harder the study of their
behaviours. Different strategies were defined in order to overcome this complexity, like the emergence of
the field of Systems of Systems (SoS) Engineering (Ackoff 1971; Boardman and Sauser 2006; Nielsen
et al. 2015). SoS break down the problem of complexity by considering that a system is a collection of
individual and heterogeneous subsystems that act together in order to reach an overall goal, enhance the
global robustness and increase the reliability of a complex monolithic system (Jamshidi 2008; Zeigler et al.
2016).
Furthermore, these systems usually need to be able to quickly adapt their behaviours to the conditions of
the executing environment. For instance, multimedia and crossmedia applications like video games need to
adapt themselves to the user profiles, in order to propose a better user experience (Streicher and Smeddinck
2016). In the case of emergency situations, systems need to adapt their actions to unexpected situations
(Whittle et al. 2009; Esfahani and Malek 2013). However, Verification and Validation of the behaviours of
such systems is harder while it is difficult to predict situations outside of the specifications. Especially in
simulation, these assumptions impose the ability to infer new scenarios respecting the initial experimental
frame, and bringing a real meaning for the system under study. In other words, verifiying and validating
self-adaptive systems imply the ability of generating dynamically new adapted and validated scenarios.
From this statement, we can make a parallel between the problems addressed by the literature of Interactive
Software and the literature of Verification, Validation and Test. While automated techniques for software
test case generation has been intensively studied for the last decades (Lu 1994; Anand et al. 2013) the
concept of self-adaptive scenarios is a main key in Digital Gaming (Rempulski et al. 2009; Chowdhury
and Katchabaw 2013; Tregel et al. 2017), in which the story tries to fit to the user decisions in order to
make it more attractive for the player. In Serious Gaming and Interactive Environment for Human Learning,
adaptability is important while it enables the deployment of different approaches that match to the needs of
each learner (Ismailović et al. 2012; El-Kechaï et al. 2015; Lavoue et al. 2018). In the case of Simulation of
Self-adaptive Software (SaS), the problem of specifying and generating relevant scenarios to verify model
adequacy and correctness is addressed (Muñoz-Fernández et al. 2015).
In short, scenario can independantly refer to stories, specifications, ordered steps or instructions, or ordered
events. As a consequence, we assume that the problem of generating adaptive scenarios for the Verification
and Validation of Simulation Models can be resolved by studying the problem from the point of view of
self-adaptive verified and validated storytelling. Indeed, storytelling refers to the system that generates a
story (i.e a scenario). Then, adaptive storytelling would be able to generate adaptive scenarios which acts
on a self-adaptive model (the system under study). Then, if the self-adaptive model is coupled to the self-
adaptive storytelling in a feedback loop, then the global model would evolve as a self-verified and validated
self-adapted system.
Therefore, we propose in this paper, as preliminary work, a way to model adaptive storytelling by exploiting
the hierarchical structure of Discrete-Event System Specifications (DEVS) in the case of Virtual Worlds and
Digital Games. Section 2 makes a recall of definitions and related works about Adaptive and Interactive
Storytelling (IS). Section 3 introduces an approach for modelling IS using DEVS. Finally, Section 4 focuses
on a example of quest generation.

2 DEFINITIONS AND RELATED WORKS

2.1 Story and Storytelling

Storytelling is a concept that have been existing since the dawn of time as a kind of suspenseful knowledge
transfer using exciting stories and immersion. From a conceptual point of view, it is a technique that sum-
marizes past experiences by introducing a correspondance between words and sequences of ordered events

359
 Yacoub, Nicolescu, Hamri, and Frydman

(Labov 1972). Andrews et al. (2009) says that a story is a mental structure of sequences of events according
to different models (Hurme 2016).
For instance, the Plot Diagram defined by Aristotle (Figure 1) is divided into three narrative blocks, orga-
nized around five phase:

• Exposition introduces the scene, the characters and the situation of the story;
• Conflict introduces the dramatic narrative element which creates exciting story and introducing prob-
lems;
• Rising action is a sequence of consequences of the conflit;
• Climax is the peak of the consequences;
• Falling actions is a sequence of actions that resolve the problems;
• Resolution is the end phase of the story.

Figure 1: Aristotle’s Plot Diagram.

Other story models follow a similar structure. Then, at a conceptual model, narrative is defined by two
elements (Figure 2): a Plot (Forster 1927; Dibell 1999) is a macro-structure that organizes the timeline of
events expressed by a story. Each component of a plot can be seen as a subplot that defines a coherent sub-
story. A story is thus a micro-structure containing a sequence of ordered events called narremes (Dorfman
1969; Wittmann 1975). This means that the sequence of events is ordered both by the Plot and by the Story
itself. Therefore, a narrative can be seen as a hierarchical structure of coherent stories.

Figure 2: Plot and Narreme.

From this definition, we can deduce that a plot is a sequence of not necessarily ordered events, just a trace,
while a story is a sequence of ordered events bound by causal relationships. Consequently, different models
of plot can be defined (Göbel et al. 2009):

360
 Yacoub, Nicolescu, Hamri, and Frydman

• linear plot generates linear stories where events are chronologically ordered. If the narrative medium
(i.e the novel, movie, etc.) has multiple narratives, each of them are independant;
• non-linear plot (Figure 3) generates stories in which the events seem to be not bound by any chrono-
logical or causal relationship. In fact, inside each narreme, the events are well-ordered, but the
subplots can be told in any orders. This means that events between subplots are not necessarily
ordered;
• branching plot models the ability to generate different events in each subplot. The global story is thus
an oriented graph, while the plot remains linear. This kind of narrative, employed in video games,
gives the feeling that different stories exist inside a unique media, or that the story has multiple ends.
• interactive plot is a plot where the storyline is not defined. Only the settings (i.e the elements of
the story like actors, characters, etc.) and possible narrative situations are described. In this kind of
storytelling, the player creates its own story. The story adapts itself to the user.

Figure 3: Non-linear and Branching Plot.

However, the main problem of interactive storytelling is the difficulty to validate the storyline, while it is
likely impossible to check all the possible stories. Furthermore, the complexity grows with the size of the
settings. Therefore, some restrictions need to be imposed in order to ensure the coherence of the generated
stories.

2.2 Interactive Storytelling

Many approaches have been explored in the literature of interactive and adaptive storytelling. Firstly, multi-
agent approaches are based on a common architecture (Arinbjarnar et al. 2009; Bostan and Marsh 2012):

• The Drama Manager is responsible for guiding the narrative by executing the best story events in a
coherent sequence and reconciling contradictory plots.
• The Agent Model handles the behaviours of the non-player characters according to the drama.
• The User Model keeps track of player choices.

361
 Yacoub, Nicolescu, Hamri, and Frydman

In Plot-based approches (Magerko 2003), the plot is seen as a state transition model. Each scene in a plot
are represented by a graph of desired states to be reached. Each state are connected using logical clauses.
In order to generate a story, the Director compares the behaviour of the user with the valid path written by
the storyteller. A predicted user behaviour model completes the user model in order to prevent errors. If an
error is detected, for instance that a path cannot be reached from a given action of the user, some decisions
are performed to change the world or to guide the user to an action that leads to a path of a the planned story.
The main limitation is that the plot must be written, meaning the system does not really adapt to the user
actions.
Rempulski et al. (2009) go further by introducing model-checking to drive and validate the storytelling. In
this approach, the plot is not a transition-state model but a set of two controlled automata: the scenario that
represents the set of the game’s entities and their possible evolutions, and a decision model that represents
the decision of theses entities. Adaptation is then made by chosen a new path in the controlled scenario
model according to the events generated by the actions of the user. Some paremeters like drama intensity,
difficulties or challenges can be thus changed.
In Character-Based approaches (Cavazza et al. 2002), the actions of each agent in the story are planned
using a Hierarchical Task Network (HTN). Each HTN represents several decompositions for the main task
given by the Drama Manager. Then, the decision making is done by two processes. On the one hand, the
graph resulting from the interleaving of all the HTNs is explored depth-first. An heuristic function evaluates
the score of each possibility to maximize the score of the goal. On the other hand, backtracking helps to
resolve narrative relevance with two mechanisms: situated reasoning and action repair (Paul et al. 2011).
Situated reasoning allows obtaining a specific resulting state in a given situation. Action repair consists of
finding a new state if the current one lead to a path that does not satisify the executability conditions. If this
methodology allows generating coherent stories, the problem is that it does not take into account intertwined
plots and dependancies between actions, which greatly increase the complexity. Storylines need thus to be
simple, and users are always seen as observers.
We propose in this paper to extend these works by using DEVS for modelling interactive storytelling. In-
deed, from the definitions given previously, a story can be seen as a hierarchical event-oriented model.
Therefore, using DEVS (Zeigler 1976; Zeigler et al. 2000) for modelling stories seems to be natural, while
simulation can help the storytelling to adapt the story progressively to the evolution of the virtual world. In
fact, we assume that it is possible to replace an entire subplot by another one using this technique as we see
in the next section. Moreover, quantitative aspect of time can be also taken into account, unlike the existing
approaches which are interested only in order of events.

3 MODELLING THE STORYTELLING USING DEVS

3.1 Story Modelling and Architecture

Our proposed architecture consists of an hybrid approach based on hierarchical layers. At the macro-level,
an approach similar to the plot-based one is applied in order to make the plot evolving according to a
storyline. At the micro-level, planning is used to reach the goal of each subplot. As a consequence, it is
possible to change the story in three manners:

1. Changing the settings. We mean replacing the actors or the objectives by equivalent ones at the
micro-level.
2. Changing the structure at the macro-level. We mean replacing a subplot by another one or dynami-
cally adding/removing subplots.
3. Changing the storyline. We mean generating new acceptable conditions of execution.

362
 Yacoub, Nicolescu, Hamri, and Frydman

The two first modifications can be achieved thanks to the DEVS architecture, while the third one can be
achieved using combined model-checking and simulation (Yacoub et al. 2017; Zeigler et al. 2017). The
structure of the story is then defined as follow.
Definition 1. We call storyline, a logical formula that the story must verify at each step. The logical formula
expresses a set of event that must be generated by the storytelling.
Definition 2. A plot is thus any sequence of events, a trace, that fulfills the storyline. A narreme is thus by
definition a plot. While the story is modelled using DEVS, a narreme is a DEVS coupled model

M =< X,Y, EIC, EOC, IC, SELECT >

where

• X is the set of input events. These events come from the external environment and models the actions
of the user;
• Y is the set of story events. The sequence of external events emitted by the system; it must match the
storyline;
• EIC and EOC represents the coupling between a subplot and its parent;
• IC represent the coupling between the narreme inside a plot.
• SELECT allows priority handling when two subplots are resolved in the same time.

Definition 3. At the micro-level, a scene is a collection of entities. An entity is a DEVS model A representing
the settings. For instance, A can be a model of a character, objects, actions, etc. As a consequence, an actor
in a scene can be instantiated multiple times in a story. Indeed, an actor can act in multiple narreme in the
same time.

While we do not enforce the model representing the entities, any formalism can be used in order to model the
actors. The only condition is that they implement the DEVS-Bus (Zeigler et al. 2000) or can be encompassed
in DEVS models.
It is important to understand the main difference with other approaches: an entity can evolve in several plots
in the same time, while a plot does not necessarily corresponds to a scene.
To illustrate this approach, consider the following narremes:

1. A - Begin of the story;

2. B - The good knight must deliver the princess;
3. C - At the same time, the bad knight must deliver the princess;
4. D - End of the story.

Consider the following requirement: "After some adventures, the princess must be delivered. She will finish
by enjoying life with its savior".
Then, we define the storyline as
A =⇒ (−) =⇒ D

where "−" represents any events.

The underlying coupled DEVS model is given in Figure 4. A, B, C and D represent scenes in which live
DEVS models of the good knight, the bad knight and the princess. The green arrows mean that when the plot

363
 Yacoub, Nicolescu, Hamri, and Frydman

A is finished, an event is emitted to active the intermediate plot, and when the intermediate plot is finished,
the final plot is enabled. This enforces the linearity of the plot.

Figure 4: Example of the Princess and its Saviour.

In the same manner, narremes B and C can have interactions between them, while the actors that they involve
can evolve in different scenes in the same time. Indeed, we mean that, if the bad knight saves the princess,
the good knight can not save her anymore for example.
Narreme B can be itself refined in a DEVS coupled model (Figure 5) representing another branching subplot:

1. BA1 - The good knight must find Excalibur;

2. BB1 - The good knight must go the castle;
3. BC1 - The good knight must kill the big dragon;
4. BD1 - The good knight must open the door of the cell.

Figure 5: Example of the good Knight and how saving the Princess.

Then, what happends if Excalibur is found by another player ? In fact, while the scene is composed by each
entity, if Excalibur is find in the narreme C, an event is also emitted in B by Excalibur itself through the
coupling. This means that BA1 is automatically cancelled, and the narreme B cannot fullfilled its storyline.
Depending on the structure of the storyline, we can then adapt the story by acting on the part that does not
respect the requirement. For example, a possible strategy would be to replace BB1 by "Kill the bad knight
which has Excalibur". Replacement can also be done using a semantic evaluation of the plot. For instance,
we can classify the entities by types and replace them with their equivalent. However, this point is out of the
scope of this paper.
This kind of architecture have a major drawback: while the actors can act independantly in any plots,
we need to instantiate them several times. This leads to an explosion of the statespace. One solution
would consist on using shared instances of DEVS model (Dalle et al. 2008). Another solution consists
of outsourcing the scenario. By uncoupling the logical model of the game, and the storytelling, we allow
reusability and more adaptability.

364
 Yacoub, Nicolescu, Hamri, and Frydman

3.2 Outsourcing the Scenario

In this approach, we consider that the storytelling and the virtual worlds are two differents systems coupled
in a feedback-loop. Then, agent model and user model are encompassed into a DEVS coupled model which
is coupled with the storytelling previously defined. Similar to the plot-based approach, when an action is
performed by the system, an event is emitted to update the story. Conversely, a change in the storytelling
emit an event to the world model for enabling or disabling elements. In short, storytelling acts more like a
monitor of the specifications of the game.
However, modifying the structure of the world for generating new contents is harder while removing an
actor would need to explore and check the entire statespace of the storytelling. This could be impossible by
considering timed events (while the story is a DEVS model, it can integrate complex timed relation between
plots, events and entities).

4 APPLICATION : THE GIRL IN THE HOUSE

In order to validate our approach, we develop a prototype of adaptive scenario in a video game. The sequence
of quests is written using a tool (Figure 6), and then translated into a DEVS coupled model. While we are
just focusing on the hierarchical structure, we do not provide the implementation of all the entities as DEVS
models (i.e. characters, items, etc.).

Figure 6: Example of Quest Writter.

In this example, the plot is divided into three rooms: the kitchen, the bedroom and the hall. The player has
the possibilty in the hall to read two messages and to navigate freely in the house. A key lays into a drawer.
One of the message enables the ability to speak to a character which asks the player to get the key in the
drawer.
Modelling the narremes gives these steps and the underlying DEVS coupled model (Figure 7):

• A1 - The Player reads the message 1.

• A2 - The Player reads the message 2.
• B1 - The Player goes to the bedroom.
• C1 - The Player speaks to the the character.
• D1 - The character asks the player to pick the key.
• E1 - The player picks the key.

365
 Yacoub, Nicolescu, Hamri, and Frydman

Figure 7: DEVS Coupled Model of the Girl in the House.

We define the storyline as :

A1 ∨ (A2 =⇒ B1)
B1 =⇒ C1
(C1 ∧ ¬key) =⇒ D1
D1 =⇒ (E1 ∧ key)

What happends if the player picks the key before speaking to the character ? In fact, readers must remember
all these plots occur in the same scene. Each entity is instantiated in each DEVS plot model. Then, when
the key is taken by the player, an event is immediately emitted to C1 at this point of the story through the
DEVS mechanism. While in C1 the condition to enable D1 is to have no key, simulation allows detecting
that the storyline would not be met, because E1 was emitted before C1. Then, when E1 is emitted, one repair
strategy consists of replacing the key by any equivalent objet in the storyline. Therefore, D1 is dynamically
replaced by D1+ : "The character asks the player to pick the key 2", and E1 by E1+: "The character picks
the key 2". This is done by just replacing the model of the key by another one in the DEVS structure of the
scene. The storyline becomes:
A1 ∨ (A2 =⇒ B1)
B1 =⇒ C1
(C1 ∧ ¬key2) =⇒ D1+
D1+ =⇒ (E1 + ∧ key2)

If we use the shared instance architecture, replacing the key involved in D1 by the key involved in D1+ also
resolves all the dependencies. The scenarios is automatically fully adapted to the new story.

5 CONCLUSION
In this paper, we propose some preliminary works to model and simulate adaptive stories in virtual worlds
using the DEVS formalism. This allows flexibility while the hierarchical structure allows splitting a sce-
nario into small units which can be semantically evaluated. Elements of the scene are modelled using atomic
DEVS model, inside each story unit. Therefore, when a path cannot be reached, a story unit can be dynam-
ically replaced by another equivalent one in order to generate a new path. Therefore, this allows the played
scenario to differ from the initial one. By coupling this methodology with a combined formalism (Yacoub
et al. 2017), validating the new scenario ensures that the story stays coherent with the inital one. Time is
also taken into account: the scenario is not just only a set of qualitative events, but also quantitative.
We also suggest that the storytelling can be considered as an independant system which is coupled with
the game in a loop: the generated scenario influences the game which in turn influences the scenario. This
allows us to consider the storytelling as an experimental frame generator for a virtual world. The logic

366
 Yacoub, Nicolescu, Hamri, and Frydman

of the story (the scenario as a set of output ordered events) is outsourced from the logic of the world (the
modelled system). In this way, we define a manner to generate dynamically a new experimental frame from
an existing one, and from the system itself. While the scenario generator is itself a DEVS model, a similar
reasonning can be applied on it.
Future works concerns the study of the complexity of the generated DEVS model. Indeed, modelling a
scenario can be considered from two different execution contexts: the model can be "plot-oriented", or
"scene-oriented". In both cases, some redundant models need to be introduced for modelling the entities
which can act in parallel plots or in parallel scenes. In these cases, the use of shared instances can lead to
some situations in which the feedback loop prohibition is not respected. Finding a way of modelling this
context in DEVS is interesting in order to reduce the size of the generated statespace. Furthurmore, we
need also study the use of Dynamic-Structure DEVS for adding and removing plots and entities. Validation
using simulation using this approach must also be addressed. Others questions concern how performing
the semantical evaluation of a story unit, and how using machine learning in order to take into account
parameters which come from outside of the system (user experience, etc.). Indeed, the replacement is
currently done by making classes of equivalent entities, which needs the definition of an ontology to keep
the meaning of a story. This adds a level of validation while the semantics can be changed over the discovery
of new scenarios.

ACKNOWLEDGMENTS
Great thanks to Pr. Bernard Zeigler for its advices about this work. Great thanks also to Delphine Routin
and Sylvain Roux for their help in the design of the case study.

REFERENCES
Ackoff, R. L. 1971. “Towards a system of systems concepts”. Management science vol. 17 (11), pp. 661–
671.
Anand, S., E. K. Burke, T. Y. Chen, J. Clark, M. B. Cohen, W. Grieskamp, M. Harman, M. J. Harrold,
P. Mcminn, A. Bertolino et al. 2013. “An orchestrated survey of methodologies for automated software
test case generation”. Journal of Systems and Software vol. 86 (8), pp. 1978–2001.
Andrews, D. H., T. D. Hull, and J. A. Donahue. 2009. “Storytelling as an instructional method: Definitions
and research questions”. Interdisciplinary Journal of Problem-based Learning vol. 3 (2), pp. 3.
Arinbjarnar, M., H. Barber, and D. Kudenko. 2009. “A critical review of interactive drama systems”. In AISB
2009 Symposium. AI & Games, Edinburgh.
Boardman, J., and B. Sauser. 2006. “System of Systems-the meaning of of”. In System of Systems Engineer-
ing, 2006 IEEE/SMC International Conference on, pp. 118–123. IEEE.
Bostan, B., and T. Marsh. 2012. “Fundamentals of interactive storytelling”. AJIT-e vol. 3 (8), pp. 19.
Cavazza, M., F. Charles, and S. J. Mead. 2002, Jul. “Character-based interactive storytelling”. IEEE Intelli-
gent Systems vol. 17 (4), pp. 17–24.
Chowdhury, M. I., and M. Katchabaw. 2013. “A Software Design Pattern Based Approach to Adaptive
Video Games”. In ADAPTIVE 2013, the Fifth International Conference on Adaptive and Self-Adaptive
Systems and Applications, pp. 40–47. Citeseer.
Dalle, O., B. P. Zeigler, and G. A. Wainer. 2008, Dec. “Extending DEVS to support multiple occurrence in
component-based simulation”. In 2008 Winter Simulation Conference, pp. 933–941.
Dibell, A. 1999. Elements of Fiction Writing - Plot. Elements of Fiction Writing. F+W Media.

367
 Yacoub, Nicolescu, Hamri, and Frydman

Dorfman, E. 1969. An Introduction to Narrative Structures. University of Toronto Press.

El-Kechaï, N., J. Melero, and J.-M. Labat. 2015, June. “Adaptation de serious games selon la stratégie
choisie par l’enseignant : approche fondée sur la Competence-based Knowledge Space Theory”. In
7ème Conférence sur les Environnements Informatiques pour l’Apprentissage Humain (EIAH 2015),
edited by C. C. D. M. e. L. O. Sébastien George, Gaëlle Molinari, pp. 294–305. Agadir, Morocco.
Conférence EIAH 2015.
Esfahani, N., and S. Malek. 2013. Uncertainty in Self-Adaptive Software Systems, pp. 214–238. Berlin,
Heidelberg, Springer Berlin Heidelberg.
Forster, E. 1927. Aspects of the Novel. RosettaBooks.
Göbel, S., F. Mehm, S. Radke, and R. Steinmetz. 2009. “80days: Adaptive digital storytelling for digital
educational games”. In Proceedings of the 2nd international workshop on Story-Telling and Educational
Games (STEG’09), Volume 498.
Hurme, Jarkko 2016. “Storytelling in Video Game : Creating a Narrative for management game”.
Ismailović, D., J. Haladjian, B. Köhler, D. Pagano, and B. Brügge. 2012. “Adaptive Serious Game Devel-
opment”. In Proceedings of the Second International Workshop on Games and Software Engineering:
Realizing User Engagement with Game Engineering Techniques, GAS ’12, pp. 23–26, IEEE Press.
Jamshidi, M. 2008, May. “System of systems engineering - New challenges for the 21st century”. IEEE
Aerospace and Electronic Systems Magazine vol. 23 (5), pp. 4–19.
Labov, W. 1972. “The Transformation of Experience in Narrative Syntax.”. [06.02.2002] Inst. f. allg. und
angewandte Sprachwissenschaft/Abt. PhASI <18/294>.
Lavoue, E., B. Monterrat, M. Desmarais, and S. George. 2018. “Adaptive Gamification for Learning Envi-
ronments”. IEEE Transactions on Learning Technologies.
Lu, P. 1994. “Test Case Generation for Specification-based Software Testing”. In Proceedings of the 1994
Conference of the Centre for Advanced Studies on Collaborative Research, CASCON ’94, IBM Press.
Magerko, B. 2003. “Building an interactive drama architecture”. In In First International Conference on
Technologies for Interactive Digital Storytelling and Entertainment.
Muñoz-Fernández, J. C., G. Tamura, I. Raicu, R. Mazo, and C. Salinesi. 2015, July. “REFAS: A PLE Ap-
proach for Simulation of Self-Adaptive Systems Requirements”. In SPLC 2015, Volume 1 of Proceed-
ings 19th International Software Product Line Conference, pp. 444. Nashville, United States, Vanderbilt
University.
Nielsen, C. B., P. G. Larsen, J. Fitzgerald, J. Woodcock, and J. Peleska. 2015, September. “Systems of Sys-
tems Engineering: Basic Concepts, Model-Based Techniques, and Research Directions”. ACM Comput.
Surv. vol. 48 (2), pp. 18:1–18:41.
Paul, R., D. Charles, M. McNeill, and D. McSherry. 2011. “Adaptive Storytelling and Story Repair in
a Dynamic Environment”. In Proceedings of the 4th International Conference on Interactive Digital
Storytelling, ICIDS’11, pp. 128–139. Berlin, Heidelberg, Springer-Verlag.
Rempulski, N., A. Prigent, P. Estraillier, V. Courboulay, and M. Perreira Da Silva. 2009, November. “Adap-
tive storytelling based on model-checking approaches”. IJIGS vol. 5 (2), pp. 33.
Streicher, A., and J. D. Smeddinck. 2016. Personalized and Adaptive Serious Games, pp. 332–377. Cham,
Springer International Publishing.
Tregel, T., J. Alef, S. Göbel, and R. Steinmetz. 2017, Oct. “Towards Multiplayer Content Online Adaptation
using Player Roles and their Interactions”. In Proceedings of The 11th European Conference on Games
Based Learning, edited by D. J. G. Dr Maja Pivec, pp. 677 – 686. Graz, Austria, Academic Conferences
and Publishing International Limited.

368
 Yacoub, Nicolescu, Hamri, and Frydman

Whittle, J., P. Sawyer, N. Bencomo, B. H. C. Cheng, and J. M. Bruel. 2009, Aug. “RELAX: Incorporating
Uncertainty into the Specification of Self-Adaptive Systems”. In 2009 17th IEEE International Require-
ments Engineering Conference, pp. 79–88.
Wittmann, H. 1975. “Théorie des narrèmes et algorithmes narratifs”. Poetics vol. 4 (1), pp. 19 – 28.
Yacoub, A., M. E. A. Hamri, C. Frydman, C. Seo, and B. P. Zeigler. 2017. “DEv-PROMELA: an extension
of PROMELA for the modelling, simulation and verification of discrete-event systems”. International
Journal of Simulation and Process Modelling vol. 12 (3-4), pp. 313–327.
Zeigler, B. P. 1976. Theory of Modeling and Simulation. John Wiley.
Zeigler, B. P., T. G. Kim, and H. Praehofer. 2000. Theory of Modeling and Simulation. 2nd ed. Academic
Press, Inc.
Zeigler, B. P., J. J. Nutaro, and C. Seo. 2017. “Combining DEVS and model-checking: concepts and tools
for integrating simulation and analysis”. International Journal of Simulation and Process Modelling vol.
12 (1), pp. 2–15.
Zeigler, B. P., H. S. Sarjoughian, R. Duboz, and J.-C. Souli. 2016. Guide to modeling and simulation of
systems of systems. Springer.

AUTHOR BIOGRAPHIES
AZNAM YACOUB is a Postdoctoral Researcher in the Computer Sciences Department of Aix-Marseille
University and Polytechnique Montreal. He holds a PhD in Computer Sciences from Aix-Marseille Univer-
sity. His research interests lie in Software Engineering, Verification and Validation, Modelling and Simula-
tion especially in Digital Gaming. His email address is aznam.yacoub@lis-lab.fr.
GABRIELA NICOLESCU obtained her PhD degree, in 2002, from INPG (Institut National Polytechnique
de Grenoble) in France, with the award for Best Thesis in Microelectronics. She has been working at Ecole
Polytechnique de Montréal (Canada) since august 2003, where she is a professor in the Computer and Soft-
ware Engineering Department. Dr. Nicolescu’s research interests are in the field of design methodologies,
programming and security for systems with advanced technologies, such as 3D multi-processor systems-
on-chip integrating liquid cooling and optical networks. She published five books and she is the author of
more than hundred articles in journals, international conferences and of book chapters. Her email address is
gabriela.nicolescu@polymtl.ca.
MAAMAR EL-AMINE HAMRI is an Associate Professor at Aix-Marseille University. He holds a Ph.D.
in Computer Sciences form Aix-Marseille University. His research interests include Modelling and Simula-
tion using DEVS and its extensions. His email address is amine.hamri@lis-lab.fr.
CLAUDIA FRYDMAN is a Professor at Aix-Marseille University. She holds a Ph.D. in Computer Sci-
ences. Her research interests include Modelling and Simulation especially using DEVS. Her email address
is claudia.frydman@lis-lab.fr.

369