AI Planning for Unique Learning Experiences:
The Time Travel Exploratory Games Approach
Oksana Arnold
1
and Klaus P. Jantke
2
1
Fachhochschule Erfurt, Germany
2
ADICOM Software, Weimar, Germany
Keywords:
Technology-enhanced Learning, Artificial Intelligence in Education, Game-based Learning, Didactic Design,
Game Design, Pedagogical Patterns, Design of Experience, Dynamic Planning, Digital Storyboarding,
Storyboard Interpretation Technology, Time Travel Prevention Games, Time Travel Exploratory Games,
Exploratory Learning, Collaborative Learning, Environmental Education.
Abstract:
All good human educators behave adaptively and treat their students individually different according to their
needs and desires. And good educators take context conditions such as disturbances from outside into account.
They react even to unforeseeable events. The more surprising a situation, the more important is the adaptivity.
In technology-enhanced learning, there is abundant evidence for the necessity of adaptive learning technology.
But how to prepare for the unforeseeable? This problem becomes even more intriguing in advanced approaches
such as, by way of illustrattion, in learning environments that allow for unusual human learner experiences like
virtual time travel. How might a learner behave when finding herself back in time in a foreign virtual world?
The authors design digital games for environmental education that enable learners to find data from the past.
The narrative is traveling back in time and exploring the past. Successful learners return with valuable findings.
But sometimes they fail. Not everyone is familiar with time travel. Preparing the exploratory digital game –
more precisely: the time travel exploratory game – for unforeseeable behavior is an involved planning task.
For this purpose, advanced technologies of Artificial Intelligence for planning in dynamic environments such
as complex industrial processes is adopted and adapted. This leads to storyboarding of learners’ experiences.
1 INTRODUCTION
The authors aim at an approach to educational game
design, in general, and to the design of games they
call Time Travel Exploratory Games, in particular.
The basic idea, still very roughly circumscribed,
is that a certain number of players, say all school-
mates of one class, play individually and undertake
virtual journeys back in time. They explore the past
and bring all their findings back to the present time.
The phase of individual exploration is followed by a
phase of collaboration. Players report about their in-
dividual time travel adventures and present each other
their findings. They are guided to an assembly of a
result that combines the varying individual findings –
the whole is more than the sum of its parts.
The crux of game design is that – unforeseeably,
but very likely–all players will have highly individual
and mutually very different time travel experiences.
They travel virtually to varying virtual locations and
to different points in virtual time. Thus, players will
act differently and will need varying amounts of time.
Players will face different obstacles and will find
their (pre)historical assets in possibly different ways.
Despite these differences, the game should enable
all players – recall they are learners – to be successful.
Furthermore, there is the need for synchronization to
allow for a phase of collaboration in which the team’s
common research result is created.
To meet all the aforementioned conditions and
goals, the game design must not determine a certain
flow of game play. Instead, it specifies a whole space
of potential behaviors and experiences. What will
happen during game play is not fully defined at design
time, but will dynamically unfold during play time.
This resembles approaches in highly complex and
dynamic environements (Arnold and Jantke, 1994a;
Arnold and Jantke, 1994b; Arnold and Jantke, 1996)
developed for the purpose of driving disturbed indus-
trial processes back into a normal mode of operation
(Arnold, 1996). Those are AI approaches to planning.
The cited approaches are adopted and adapted.
What is called a plan in books like (Arnold, 1996)
is now a digital storyboard that unfolds at play time.
124
Arnold, O. and Jantke, K.
AI Planning for Unique Learning Experiences: The Time Travel Exploratory Games Approach.
DOI: 10.5220/0010453001240132
In Proceedings of the 13th International Conference on Computer Supported Education (CSEDU 2021) - Volume 1, pages 124-132
ISBN: 978-989-758-502-9
Copyright
c
 2021 by SCITEPRESS – Science and Technology Publications, Lda. All rights reserved

Figure 1: The permanently grounded aircraft IL-18 that is hosting “The Flying Classroom” project with the present game.
2 EMBEDDING APPLICATION &
IMPLEMENTATION
“The Flying Classroom” is the name of a very recent
edutainment project
1
that shall be equipped with a se-
ries of digital learning games.
The authors’ present contribution, on purpose,
does not deal with the issues of implementation and
application. Instead, the didactic design approach of
dynamic planning and storyboarding is in focus.
The present section is only intended, so to speak,
to embed the authors’ work into a wider context, to
see it almost literally from a bird’s eye view, and to
allow for an appraisal of its reach.
The Flying Classroom is located in a permanently
grounded Russian-made aircraft of type IL-18 (fig. 1)
of the former East-German airline Interflug.
Figure 2: Ocean warming data from 1960 ©Cheng, 2020.
1
This article from the main Thuringian newspapers
is at the time of running the CSEDU 2021 confer-
ence less than half a year old: https://www.thueringer-
allgemeine.de/regionen/erfurt/vorfreude-auf-das-fliegende-
klassenzimmer-in-erfurt-id231055542.html
Classes of pupils may visit the aircraft. According
to their age and grade of school, they get offered to
play a digital game inside the aircraft. The framing
story of every game is a virtual journey in the IL-18’s
Flying Classroom.
Every game has its particular learning content.
The present contribution reports about the first of
these games. The content is the ocean warming which
is a very intriguing issue due to the ocean currents and
the exchange of heat between ocean and atmosphere
(Rhein et al., 2013).
Investigating the ocean warming requires observa-
tions over a longer span of time and taking the ocean
world-wide into account. This is very demanding and
requires a literally global approach.
Virtually flying around the world takes care of the
problem of world-wide observations. But what about
the problem of observations over a longer period of
time? This problem is resolved by virtual time travel.
Figure 3: Ocean warming data from 2019 ©Cheng, 2020.
Players on a virtual journey back in time may take
thermal pictures like those in the figures 2 and 3.
Finally, players discuss their individual findings
AI Planning for Unique Learning Experiences: The Time Travel Exploratory Games Approach
125

and assemble a video illustrating the ocean warming
over a period of 40 years.
3 TIME TRAVEL GAMES
Time Travel is known in commercial games since
DAY OF THE TENTACLE, 1993, and SHADOWS OF
DESTINY, 2001. Na
¨
ıve variants of time travel occur
in games such as BRAID, 2008.
Furthermore, the first introduction of save points
in games, pioneered in THE LEGEND OF ZELDA,
1986, may be seen as rudimentary time travel.
3.1 Time Travel Prevention Games
To the present authors’ very best knowledge, the term
Time Travel Prevention Games has been coined on
the German Prevention Day, Frankfurt/M., Germany,
2015. It is listet there as a keyword
2
.
In a sense, SHADOWS OF DESTINY is a prototyp-
ical time travel prevention game.
In a time travel prevention game, players con-
fronted with undesired events may get the opportunity
of travelling back in time to change the past such that
the undesired event may be avoided the next time.
Concepts of time travel prevention games have
been developed and studied for varying purposes such
as crime prevention (Winter and Jantke, 2014).
They have enormous potential for health care and
for the training of accident prevention.
3.2 Time Travel Exploratory Games
The term Time Travel Exploratory Games is coined in
the present publication.
In such a game, the purpose of time travel is not
to change the past, but to explore it. To some extent,
the exploration is an important aspect of time travel
prevention games as well. Players need to explore the
virtual past for finding a way to change it as desired.
But in time travel exploratory games, exploration is
primary.
What players may bring with them when returning
to the present are experience, insights, and artefacts.
So, first of all, time travel exploration games are – by
nature – learning games.
In dependence on the subject of the study, it may
be more or less desirable or even necessary, to visit
different locations in different periods of time.
2
https://www.praeventionstag.de/nano.cms/vortraege/begriff/Time-
Travel-Prevention-Games?sb=Time+Travel+Prevention+Games
The necessity of explorations comprehensive in
space and time suggests the engagement of multiple
players. By nature, time travel exploratory games
tend to be effective multiplayer games.
Time travel exploratory games are useful for com-
petitive and collaborative learning and both at once.
4 ENVIRONMENTAL
EDUCATION
4.1 Environmental Education in
General
All big themes of environmental education are related
to changes of the world such as the ozone depletion,
the deforestation of rainforest, and the glacier melt.
Other serious problems range from the decline in the
insect population to the impending extinction of the
Javan rhinoceros.
Environmental education depends on data that
change over time and on the comparison of these data.
First, the interpretation of findings is intriguing
and especially young learners very much depend on
visualization.
Second, the study of large amounts of data may
easily become boring.
In the light of these difficulties, the two authors
propose time travel exploratory games as an approach
to environmental education.
Finding data through an adventurous process of
time travel to and exploration in the virtual past leads
to a certain bonding. Players value their individual
findings and show more interest in comparison and
interpretation. Hence, they look at visualizations of
their own findings with a particular interest.
Larger amounts of data result from the individual
explorations of members of a larger team like a class.
At this point, it is time to change the didactic method.
Individual competitive exploration is followed by
the collaborative creation of the team‘s overall result.
The following four chapters 5 6, 7, and 8 step by step
present the educational and game design. The authors
see this as a guideline for applications of their story-
boarding approach.
4.2 Ocean Warming in Particular
The article (Chen et al., 2020) is a highly topical and
alarming knowledge source demonstrating the gravity
of the situation. The quite intriguing interference of
cyclones and ocean warming is discussed in depth and
CSEDU 2021 - 13th International Conference on Computer Supported Education
126

detail by (Trenberth et al., 2018). The ocean warm-
ing invigorates tropical cyclones. On the other hand,
the strong winds of tropical cyclones keep the ocean
cooler by causing stronger evaporation.
To begin with, environmental education shall in-
form the learners about the phenomenon of ocean
warming making it intelligible both in its gravity and
in its diversity. The complexity of the problem shall
become discernible resulting in an appreciation for
collaborative studies and big data management.
5 TIME TRAVEL GAMES FOR
EXPLORATORY STUDIES OF
OCEAN WARMING INCL.
COLLABORATIVE LEARNING
The key idea as already introduced above consists in
a multiplayer digital game installed inside an aircraft.
A group of players engage in game play in which
• players undertake independently of each other a
virtual journey back in time,
• explore the past virtually,
• find artefacts that carry information about ocean
warming,
• bring these artefacts back into the present time,
• compare, discuss, and interpret their individual
findings,
• and collaborate toward a common presentation of
the aggregate of their findings.
This leads to a design of interaction, experience, game
play, and learning as summarized by the graph on dis-
play in figure 4.
Figure 4: Top level storyboard graph.
In a storyboard graph like the one in figure 4, the
nodes are called episodes. Other nodes called scenes
will be introduced later. Episodes are placeholders for
more detailled descriptions of anticipated experience.
In slightly more technical terms, the episode nodes
determine places for graph substitution. Replacement
of episodes by other graphs of the storyboard takes
place at execution time. The discussion of these tech-
nicalities is postponed.
Edges from one node to the other specify the flow
of interaction. Every edge is annotated by a logical
condition of execution. In this way, it is determined
in which conditions game play may proceed from one
episode to the other. The logical conditions seen as
formulas contain variables. These variables refer to
environmental data – game play may depend on the
day time and even on the outside wheather conditions
– and on data from the user/learner/player model in-
cluding the interaction history.
For simplicity, one may assume that the conditions
[1], . . . , [4] in figure 4 have constantly the value true.
The design of interaction, game play, and learning
proceeds via a step by step introduction of storyboard
graphs admissible for the expansion of episodes. This
is the collaborative process of storyboarding that may
proceed bottom-up, top-down, or even both at once.
Storyboarding means the organization of experience
(see (Jantke and Knauf, 2005). p. 25).
This usually results in alternative concepts of
graph expansion, i.e. varying proposals of what to
substitute for a particular episode node. Every graph
Figure 5: Storyboard graph to expand GAME PLAY.
is annotated by its substitution condition. This allows
for the design of alternatives that will be invoked in
different conditions.
By way of illustration, the episode node named in
figure 4 may be substituted by the graph on display in
figure 5.
Let us briefly describe the intentions behind the
graph to be substituted. First, the flying classroom
aircraft virtually arrives at some destination, say the
great barrier reef. Second, players get some visual
information and identify their current location. Third,
they get the opportunity of individual time travel back
in time. When they arrive in the past, fourth, play-
ers explore the environment and using the virtual in-
frared camera take a picture. Fifth, this picture is in-
serted into the infrared ocean map they have. Sixth,
all the players return individually to the present time
and, seventh, the virtual journey continues to another
location.
AI Planning for Unique Learning Experiences: The Time Travel Exploratory Games Approach
127

The complanate nodes of the graph in figure 5 are
scenes. Those have a some fixed semantics in the
game mechanics such as a scripted scene, a video, or
an audio file with a spoken text.
There is a branching point at the scene FLIGHT
CONT’D. The number of time travels may be limited
in advance or depend on the duration of game play so
far. This is controlled by the conditions [7] and [8].
After the end of this exploratory episode it follows
the episode SYNTHESIS of collaborative learning.
6 DYNAMIC PLANNING AND
DIGITAL STORYBOARDING
Given any digital storyboard as a finite collection of
directed graphs as on display in the figures 4 and 5,
what players/learners will experience in the course of
interaction is largely unforeseeable due to the graph
expansion at execution time and due to the dynami-
cally changing validity of branching conditions.
There is a similarity to the treatment of seriously
disturbed complex technical processes. For example,
think of a chemical reactor in which occurs an un-
expectedly high pressure. This may be caused by an
incorrect percentage of certain chemical ingrediens.
But if the disturbance is unexpected, nobody knows
exactly about the situation inside the reactor. There
are larger numbers of measures to undertake, but the
effects are not fully foreseeable. In dependence on the
dynamic data, the process of driving the system back
into a normal mode of operation unfolds during inter-
action. It is not planned in advance. Domain experts
know about possible effects of all the measures they
have. And they know about conditions of execution.
The measures in such a technical setting resemble
the graphs as illustrated in the preceding section 5.
Based on this similarity, the authors’ approach is
to adopt and adapt the AI planning concepts from
(Arnold, 1996) and to transform them into concepts
of storyboarding time travel exploratory games.
6.1 Essentials of AI Dynamic Planning
Because the technicalities such as precise syntax and
implementation details remain, so to speak, under the
hood, an intuitive and widely informal presentation is
appropriate.
As introduced above, directed graphs are the
key concept for representing anticipated (inter)action.
There is, more precisely, the concept of pin graphs.
Every such graph has entrance nodes and exit nodes.
In the technical planning domain, the set of all nodes
of a pin graph is partitioned into two classes. Some of
the nodes are called compound nodes. All the other
nodes are called atomic nodes. Atomic nodes have
a meaning in the technical domain such as a human
intervention or a program to be executed. Compound
nodes may be subject to later substitution.
A plan is a finite hierarchically structured family
F = {G
i
}
i∈I
of pin graphs, where I is any index set.
For simplicity, one may imagine I to contain positive
natural numbers, perferably without gaps inbetween,
i.e. I = {1, 2, 3, . . . , k}. k is the exact number of pin
graphs in the plan F . In practice, it is advantageous
to use more expressive names instead of numbers.
One graph describes on the most general level and
with most rough granularity the overall plan. This is
the top-level graph that may be named G
1
in case the
index set consists of positive natural numbers.
By way of illustration, recall our graph on display
in figure 4. This is a top-level graph, as we put it
above, on the most general level and with most rough
granularity.
Inside every graph, logical conditions are used to
annotate edges. These conditions determine whether
or not an edge may be used to step forward from one
node to another at execution time.
Inside every graph, compound nodes get assigned
graphs (for simplicity, just the indices of graphs) to
determine which graph may be taken as a substitute
for the node. Recursive substitutions are admissible.
Outside every graph, so to speak, there exists an
annotation by a logical condition of deployment.
The regulations of graph substitution establish the
hierarchical structure of the plan F .
Planning is the process – team-based, possibly
distributed over space and time, alternatively bottom-
up or top-down or both at one – of constructing all the
graphs of F including the internal annotations of the
edges that determine the conditions of the usage of
edges as well as the external substitution conditions
that control graph substitution for compount nodes.
Last but not least, planning means to assign to every
atomic node an operational semantics in the domain.
The unrivaled modularity of the approach bears its
unprecedented flexibility (Arnold, 1996).
The logical conditions contain variables of which,
in the technical domain, many are numerical. At the
time of planning, most values at some future points of
time are unknown.
When some disturbance occurred, plan execution
begins with G
1
. The atomic nodes are executed,
dynamically emerging knowledge enables decisions,
substitutions take place, the plan unfolds.
CSEDU 2021 - 13th International Conference on Computer Supported Education
128

Figure 6: Dynamic knowledge relevant to plan execution.
6.2 Digital Storyboarding as a Form of
AI-based Dynamic Planning
Many industrial processes are highly complex and
disturbed industrial processes may easily become
very dangerous. However, every human being is
much more complex than any chemical installation.
Consequently, every effort to design adaptivity for the
purpose of meeting human needs and desires shall be,
at least, as powerful as the best AI approaches that
adapt to technical necessities.
The AI planning approach of (Arnold, 1996) is
carried over to planning human learners’ experiences,
especially to the design of time travel exploratory
games.
A storyboard is a finite hierarchically structured
family of pin graphs. Internally and externally, resp.,
the graphs are annotated in the same way as described
in the preceding section 6.1.
Nodes of a graph that are scenes have an opera-
tional semantics in the time travel exploratory game.
By way of illustration, see the scene of the graph in
figure 5 that is named ARRIVAL AT DEST. This scene
may be implemented by an audio file in which the vir-
tual aircraft captain announces the arrival – virtually –
at a certain destination. Further implementations are
possible such as a video showing the aircraft captain
during his announcement or just a textual information
on the screen.
Every storyboard has exactly one graph on the top
level 1 like the graph on display in figure 4. Graphs
that may be substituted for an episode of the graph
on level 1 are said to be on level 2. In general, given
a particular graph G
j
and n being the highest level
on which a graph G
i
exists containing an episode that
may be substituted by G
j
, this graph is said to be on
level n+1. The lower the level, i.e., the larger the level
number, the finer the granularity of interaction design.
This corresponds to layered languages of ludology
as introduced, studied, and applied a good decade ago
in (Jantke, 2006) and (Lenerz, 2009).
In a sense, digital storyboards are operational.
They may be seen as computer programs of the inter-
active media specified. A complete digital storyboard
of a time travel exploratory game defines the game’s
implementation. The crux is that the program can
not be compiled to executable code, because dynamic
data decisive for the evaluation of logical conditions
in the storyboard are not available prior to execution.
Alternatively, the storyboard may be interpreted at
play time (Fujima et al., 2013). In dependence on the
dynamics, interpretation unfolds the one or the other
experience of game play and learning. Interpretation
is adaptive to human needs and to varying context.
7 DIDACTIC AND GAME DESIGN
Dovetailed didactic design and game design take
place as digital storyboarding. The result is a highly
modular digital document describing not only one
anticipated flow of human-system interaction, but a
potentially infinite space of human game play and
learning experiences.
The digital storyboard as a collection of annotated
graphs has both a rather appealing and intuitively
comprehensible appearance and the syntactic details
of a computer program ready for interpretation.
Storyboard graphs are intentionally small to keep
the overview and to be able to discuss the smallest
details of pedagogy and of ludology.
Interdisciplinary designer teams of experts such
as educators, ludologists, AR and VR specialists and
others may maintain even contradictory alternatives.
Decisions may be postponed during the design phase
and some may even be taken as late as at execution
time, i.e., when the digital game is running, when fun,
excitement, and learning interfere.
Higher level design decisions are represented in
higher level graphs. By way of illustration, the order
of the episodes GAME PLAY and SYNTHESIS in the
top level graph of figure 4 reflects the basic decision
that exploration precedes collaboration.
On a slightly lower level, one has to specify how
to individually travel back in time; an expansion of
the corresponding episode in figure 5. Several of the
ideas are reflected by the storyboard graph on display.
All nodes of this graph are scenes to be implemented.
According to the overarching game design idea
that individual explorations are followed by a phase of
collaboration in which different players bring in their
mutually different (!) findings, there is developed the
following game mechanics. Players are encouraged
AI Planning for Unique Learning Experiences: The Time Travel Exploratory Games Approach
129

Figure 7: Individual time travel back in time.
to go on a journey back in time. If this is a player’s
first time travel (condition [1]), there appears a list of
target time periods to selct one. We have 240 alterna-
tives resulting from 60 years times 4 quartes of data.
On the occasion of a later time travel (condition [2]),
players are guide to the target time selected earlier.
8 GAME DESIGN PATTERNS
AND DIDACTIC PATTERNS
The pattern concept became explicit in science and
technology by Christopher Alexander’s related work
in architecture (Alexander, 1979).
At almost the same time, Dana Angluin developed
a very precise and lucid approach to patterns and their
instances (Angluin, 1980). Trendsetting is Angluin’s
insight that there might be patterns in science and
technology, but what we can perceive are only their
instances. This leads directly to the pattern inference
problem. Seeing a few or, perhaps, a large amount of
instances, what is the pattern behind? This problem is
usually an intriguing one.
In some areas, the pattern concept is rather vague
and the borderline between patterns and instances gets
blurred (Pedagogical Patterns Advisory Board, 2012).
Surprisingly, despite the appealing title of the
book (Bj
¨
ork and Holopainen, 2004), this work does
not contribute much to a dovetailed didactic and game
design, as many of the concepts discussed are far from
being patterns.
That patterns are a key concept of game-based
learning is illustrated by (Jantke, 2012) where the oc-
currences of patterns – more correctly, of instances of
patterns – may be interpreted as indicators of mastery.
Thus, patterns and instances are key to assessment.
8.1 From Patterns to Instances
Let us see the pattern inference problem the other way
around. When a pattern is designed, this determines
all its possible instances.
Human players and learners do not experience the
abstract concepts behind some game design, but what
really happens. In other words, what designers really
want are the instances, not the patterns.
The usage of design patterns means the usage of
general principles that are likely to result in concrete
interactions – the instances – that, hopefully, have an
intended impact.
Among the game design patterns there are those
that appear boring, at a first glance, but are relevant to
the game mechanics. Many of these features are not
recognized by the player, because their instances run
in the background. This is mentioned here, because
those patterns must not be overlooked. For brevity,
we discuss just one: book keeping.
Whenever players make important decisions of
game play, there is the necessity to record this for
the purpose of later adaptivity. Data are send to the
server and form the game play history as well as the
human player profile. The corresponding scene is
named BOOK KEEPING (see figure 7).
One important insights is, that patterns appear as
structural properties of storyboard graphs. What does
not show structurally, that is not used semantically.
This explicates the advantage of this design approach.
Finally, there will be discussed a design pattern
that is interesting, because it may be seen as a didactic
pattern and a game design pattern as well: self-reliant
interaction.
Figure 8: Onboarding incl. time travel.
Onboarding is a terminus technicus of the digital
games area. It denotes an early phase of game play
(see figure 4) in which players learn how to play and
what to do without the need to read any manual or
anything like this. The usage of onboarding is a game
design pattern.
The players get an introduction into time travel
and how to use the virtual IR camera. They make their
first IR shot and get help in interpreting the value of
their finding. After that, they are encouraged to try it
again self-reliantly (see figure 8).
This is the basis for future self-determined interac-
tion during the exploratory phase of game play, hence,
a didactic pattern of self-determination.
CSEDU 2021 - 13th International Conference on Computer Supported Education
130

8.2 From Instances to Patterns
When a time travel exploratory game is in use,
one may easily record game playing. Conceptually,
recorded game play is a string of (inter)actions that
took place. In those string, one may find instances of
what happened (Jantke, 2012).
A systematic analysis will reveal which of the
game design and didactic patterns have resulted in
anticipated human experience.
For a very rough illustration, it might happen that
during graph expansion a particular storyboard graph
is never invoked. Phenomena like that are studied in
(Knauf et al., 2010).
To say it in terms of the pattern inference problem,
observing game playing behavior that takes place,
identifying potential instances of potential patterns,
and learning the patterns behind gives feedback about
the effectivity of the principles incorporated.
9 SUMMARY & CONCLUSIONS
After coining the term time travel exploratory games,
it is the authors’ foremost intention to advocate digital
storyboarding as a methodology of truly dovetailed
pedagogical and game design – team-based, possibly
distributed over space and time, alternatively bottom-
up or top-down or both at once. It clarifies details
of didactics (Jantke, 2013) and works for large-scale
applications (Arnold et al., 2013).
ACKNOWLEDGEMENTS
The authors gratefully acknowledge the co-operation
with Steffen Avemarg and Winfried Wehrstedt within
the Flying Classroom project.
REFERENCES
Alexander, C. (1979). A The Timeless Way of Building. New
York: Oxford University Press.
Angluin, D. (1980). Finding patterns common to a set of
strings. J. Computer and Systems Science, 21:46–62.
Arnold, O. (1996). Die Therapiesteuerungskomponente
einer wissensbasierten Systemarchitektur f
¨
ur Auf-
gaben der Prozeßf
¨
uhrung, volume 130 of DISKI. St.
Augustin: infix.
Arnold, O. and Jantke, K. P. (1994a). Therapy plan genera-
tion as program synthesis. In Arikawa, S. and Jantke,
K. P., editors, Algorithmic Learning Theory, Proc. 4th
International Workshop on Analogical and Inductive
Inference (AII’94) and the 5th International Workshop
on Algorithmic Learning Theory (ALT’94), October
10–15, 1994, Reinhardsbrunn Castle, Germany, vol-
ume 872, pages 40–55. Springer–Verlag.
Arnold, O. and Jantke, K. P. (1994b). Therapy plans as hier-
archically structured graphs. WISCON Report 02/94,
HTWK Leipzig (FH), FB Informatik, Mathematik &
Naturwissenschaften.
Arnold, O. and Jantke, K. P. (1996). Representing and
processing dynamic knowledge in complex dynamic
systems. In Alpaydın, E., C¸ ilingiro
˘
glu, U., G
¨
urgen,
F., and G
¨
uzelis¸, C., editors, Fifth Turkish Sympo-
sium on Artificial Intelligence and Neural Networks,
TAINN’96, Istanbul, June 27–28, 1996, pages 351–
356. Bo
˘
gazic¸i University.
Arnold, S., Fujima, J., and Jantke, K. P. (2013). Story-
boarding serious games for large-scale training appli-
cations. In Foley, O., Restivo, M. T., Uhomoibhi, J.,
and Helfert, M., editors, Proceedings of the 5th Inter-
national Conference on Computer Supported Educa-
tion, CSEDU 2013, Aachen, Germany, May 6-8, 2013,
pages 651–655.
Bj
¨
ork, S. and Holopainen, J. (2004). Patterns in Game De-
sign. Hingham, MA, USA: Charles River Media.
Chen, L., Abraham, J., Zhu, J., Trenberth, K. E., Fasullo, J.,
Boyer, T., Locarnini, R., Zhang, B., Yu, F., Wan, L.,
Chen, X., Song, X., Liu, Y., and Mann, M. E. (2020).
Record-setting ocean warmth continued in 2019. Ad-
vances in Atmospheric Sciences, 37:137–142.
Fujima, J., Jantke, K. P., and Arnold, S. (2013). Digital
game playing as storyboard interpretation. In Pro-
ceedings of the 5th International Games Innovation
Conference (IGIC), Vancouver, BC, Canada, Sept. 23-
25, 2013, pages 64–71. IEEE Consumer Electronics
Society.
Jantke, K. P. (2006). Layered Languages of Ludology: The
Core Approach. Diskussionsbeitr
¨
age 25, TUI IfMK.
Jantke, K. P. (2012). Patterns of game playing behav-
ior as indicators of mastery. In Ifenthaler, D., Es-
eryel, D., and Ge, X., editors, Assessment in Game-
Based Learning: Foundations, Innovations, and Per-
spectives, pages 85–103. New York, Heidelberg, Do-
drecht, London: Springer.
Jantke, K. P. (2013). Pedagogical patterns and didac-
tic memes for memetic design by educational story-
boarding. In Arnold, O., Spickermann, W., Spyratos,
N., and Tanaka, Y., editors, Webble Technology,
First Webble World Summit, WWS 2013, Erfurt, Ger-
many, June 2013, volume 372 of Communications in
Computer and Information Science, pages 143–154.
Springer.
Jantke, K. P. and Knauf, R. (2005). Didactic design through
storyboarding: Standard concepts for standard tools.
In in: First International Workshop on Dissemination
of E-Learning Technologies and Applications, DELTA
2005, Cape Town, South Africa, January 3–6, 2005,
pages 20–25. Computer Science Press, Trinity College
Dublin, Ireland.
Knauf, R., Sakurai, Y., Tsuruta, S., and Jantke, K. P. (2010).
Modeling didactic knowledge by storyboarding. J.
Educational Computing Research, 42(4):355–383.
AI Planning for Unique Learning Experiences: The Time Travel Exploratory Games Approach
131

Lenerz, C. (2009). Layered languages of ludology. In
Beyer, A. and Kreuzberger, G., editors, Digitale
Spiele – Herausforderung und Chance, pages 39–52.
Boizenburg: VWH.
Pedagogical Patterns Advisory Board, editor (2012). Ped-
agogical Patterns: Advice for Educators. Joseph
Bergin Software Tools.
Rhein, M., Rintoul, S. R., Aoki, S., Campos, E., Chambers,
D., Feely, R. A., Gulev, S., Johnson, G. C., Josey,
S. A., Kostianoy, A., Mauritzen, C., Roemmich, D.,
Talley, L. D., and Wang, F. (2013). 2013: Observa-
tions: Ocean. In Climate Change 2013: The Physical
Science Basis. Cambrige, U.K., and New York, NY,
USA: Cambridge University Press.
Trenberth, K. E., Cheng, L., Jacobs, P., Zhang, Y., and Fa-
sullo, J. (2018). Hurricane Harvey links to ocean heat
content and climate change adaptation. Earth’s Fu-
ture, 6:730–744.
Winter, J. and Jantke, K. P. (2014). Formal concepts and
methods fostering creative thinking in digital game
design. In 3rd Global Conference on Consumer Elec-
tronics (GCCE 2014), Oct. 7-10, 2014, Makuhari
Messe, Tokyo, Japan, pages 483–487. IEEE.
CSEDU 2021 - 13th International Conference on Computer Supported Education
132