Models and Concepts for Socio-technical Complex Systems: Towards Fractal Social Organizations

SYSTEMS RESEARCH AND BEHAVIORAL SCIENCE
Syst. Res. Behav. Sci. 2013; 00:1–24
Models and Concepts for Socio-technical Complex Systems:
Towards Fractal Social Organizations
Vincenzo De Florio1
, Mohamed Bakhouya2
, Antonio Coronato3
,
and Giovanna Di Marzo4
1Performance Analysis of Telecommunication Systems research group
Universiteit Antwerpen and iMinds Interdisciplinary institute for Technology
Middelheimlaan 1, 2020 Antwerp, Belgium.
vincenzo.deflorio@uantwerpen.be
2School of Engineering, Aalto University
Otakaari 4, 00076 Aalto, Helsinki, Finland.
mohamed.bakhouya@aalto.fi
3Institute for High Performance Computing and Networking
National Research Council
Via P. Castellino, 111, 80131 Naples, Italy.
antonio.coronato@na.icar.cnr.it
4Institute of Services Science, Université Genève
Battelle–Bâtiment A, Rte de Drize 7, 1227 Carouge, Switzerland.
giovanna.dimarzo@unige.ch
SUMMARY
We introduce fractal social organizations—a novel class of socio-technical complex systems characterized
by a distributed, bio-inspired, hierarchical architecture. Based on a same building block that is recursively
applied at different layers, said systems provide a homogeneous way to model collective behaviors of
different complexity and scale. Key concepts and principles are enunciated by means of a case study and a
simple formalism. As preliminary evidence of the adequacy of the assumptions underlying our systems here
we define and study an algebraic model for a simple class of social organizations. We show how despite its
generic formulation, geometric representations of said model exhibit the spontaneous emergence of complex
hierarchical and modular patterns characterized by structured addition of complexity and fractal nature—
which closely correspond to the distinctive architectural traits of our fractal social organizations. Some
reflections on the significance of these results and a view to the next steps of our research conclude this
contribution. Copyright c 2013 John Wiley & Sons, Ltd.
Received ...
KEY WORDS: Socio-technical systems; Social organizations; Collective adaptive systems; Distributed
organizations; Fractal organizations
1. INTRODUCTION AND RATIONALE
Societal and technological progress have brought to a widespread diffusion of Internet-backed
computer services characterized by an ever increasing complexity, pervasiveness, and social
meaning (Fleisch, 2010). Ever more often people make use of and are surrounded by devices
enabling the rapid establishment of communication channels on top of which people may socialize
and collaborate, supply or demand services, query and provide knowledge as it had never been
possible before. Such devices become our knowledge and social backbone by means of which
∗Correspondence to: Vincenzo De Florio, Middelheimlaan 1, 2020 Antwerp, Belgium.
Copyright c 2013 John Wiley & Sons, Ltd.
Prepared using speauthmodified.cls [Adapted by the authors from speauth.cls version 2010/05/13 v3.00]

2 V. DE FLORIO ET AL.
new spontaneous forms of socialization and novel forms of self-organization may arise—as
demonstrated by knowledge ecosystems (Bray et al., 2008), cyber-physical societies (Zhuge, 2010),
and mutual assistance communities (Sun et al., 2006; Sun et al., 2010). The virtually instantaneous
diffusion of knowledge and awareness granted by such digital ecosystems is already producing
novel forms of collective intelligence and social interaction, with tremendous significance in terms
of marketing, economy, welfare, and social values (Pór, 2000).
The above mentioned novel forms of social organization and interaction are a promise of new
wealth and quality for all; at the same time, they also constitute a covenant of new solutions against
the many new problems our societies are experiencing. The overwhelming increase in the human
population and in particular of the elderly; the dissipation of valuable resources such as water,
energy, and clean air; the ill-considered management of waste are but a few examples that show
how the current social organizations are proving to be ineffective and unable to scale to the sizes
of our new “big world” (Hardin, 1968; Barabasi et al., 2013). Assistance of the elderly population
is a typical case in point: The share of the total population older than 65 is constantly increasing
worldwide (Anonymous, 2011; Anonymous, 2012), while the current organizations still provide
assistance in an inefficient and inflexible way. Though effective when the context was different
and a large amount of resources were available to manage a smaller demand, this approach is now
becoming too expensive and thus unacceptable. Merely expanding the current organizations without
properly restructuring them is simply not working anymore (De Florio & Blondia, 2008).
Promising solutions to this problem come from the domain of control systems. Such systems
have been traditionally crafted by using paradigms such as the centralized, the hierarchical, or the
heterarchical (Dilts et al., 1991). More recently additional classes of distributed control mechanisms
have been introduced so as to enhance efficiency and reduce the bottlenecks characterizing the
classical paradigms. The terms used in literature to refer to these classes is bionic, holonic,
and fractal organizations (Tharumarajah et al., 1996; Ryu, 2003). Said solutions have been
successfully applied in several domains in order to achieve “smarter organizations”, i.e. systems
characterized by greater scalability, robustness, and manageability with respect to their traditional
counterparts (Warnecke & Hüser, 1993; Tharumarajah et al., 1998; Ryu, 2003). Our conjecture and
major starting point in the current discussion is that the above distributed mechanisms may provide
us with promising paradigms for the design of socio-technical complex systems able to make full
use of the potential inherent in our societies so as to turn them into abundant sources of valuable
assets, complex collaborative behaviors, and massive redundancy. A major challenge then becomes
being able to master the complexity of said mechanisms and create models for the organization and
the management of massively distributed open socio-technical systems based on their paradigms.
1.1. Contributions and Structure
Section 2 presents our first contribution in the above stated framework of problems, namely the
definition of a class of socio-technical complex systems based on a distributed, bio-inspired,
hierarchical organization. Said organization is characterized by a same building block that is
repeated at different layers so as to manage both self-sufficient and inter-community collective
behaviors. Its building block, called service-oriented community, is “simultaneously a part and a
whole, a container and a contained, a controller and a controlled” (Sousa et al., 2000) in that it
represents the canon (Ryu, 2003) of a fractal organization (we refer the reader to Sect. 4.1 for more
information on fractal organizations). This fact led to the name of “fractal social organizations” for
systems compliant to our models (De Florio et al., 2012).
A second contribution of this paper is reported in Sect. 3. In that section we describe a model of the
collective behaviors within non-hierarchical and static socio-technical systems. A dynamic system is
introduced to express the life-cycle of active behaviors within said socio-technical systems. Despite
so simple a formulation, the resulting dynamics produce complex emerging properties. In particular,
• The geometrical representations of the orbits of the dynamic system exhibit a spontaneous
emergence of self-similar patterns, some of which likely to be characterized by a fractal
dimension.
Copyright c 2013 John Wiley & Sons, Ltd. Syst. Res. Behav. Sci. (2013)
Prepared using speauthmodified.cls DOI: 10.1002/srbs

TOWARDS FRACTAL SOCIAL ORGANIZATIONS 3
• A structured addition of complexity appears to emerge as complex structures reveal to be the
result of a regular composition of “prime” (i.e., no further decomposable) building blocks.
• A hierarchical organization spontaneously shapes up.
It is worth remarking how several of the above properties closely correspond to the distinctive traits
of our fractal social organizations. We conjecture that so pronounced a similarity may hint at the
emergence of scalability and robustness in future socio-technical complex systems designed after
our fractal social organizations.
This article is concluded by a discussion of related concepts and disciplines in Sect. 4 and by a
summary of our preliminary results as well as our plans for future research in Sect. 5.
2. FRACTAL SOCIAL ORGANIZATIONS
The starting point of our discussion is the definition of social organization provided by Boulding
in his classic General Systems Theory article (Boulding, 1956). In the cited renowned contribution
Boulding provides a classification of system behaviors and—to the best of our knowledge—he is
the first scholar to highlight explicitly the role and significance of socio-technical complex systems
and their behaviors. Social organizations correspond to said systems and are classified by Boulding
as the most complex system category. His definition for social organizations is “a set of roles tied
together with channels of communication”. We observe how even so concise a definition already
captures several important aspects of the dynamics of collective behavior:
A set...: Social organizations are selections of societal constituents.
...of roles...: Said selections are characterized not by the identities of their constituents, but rather
by their role: Quoting again from (Boulding, 1956), in social organisations “the unit [...] is
not perhaps the person but the role—that part of the person which is concerned with the
organisation or situation in question”. In algebraic terms this translates in modeling social
organizations as multisets of roles. One such model is introduced in Sect. 3.
...tied together...: Shared goals and common opportunities may lead individuals to join forces
and constitute a new “social” entity able to exert purposeful collective behavior. We refer
to such entity as a community. Reasons for individuals to tie together into a community
may be temporary or permanent; moreover, they may be the result of purposeful active
individual behaviors (Rosenblueth et al., 1943) as a response to the onset of an environmental
condition—for instance, a threat to survival or an opportunity for economical profit—or
simpler forms of individual behavior such as those typical of animal instinct. Moreover, the
resulting social entity may be a temporary or permanent one. All these factors influence
the dynamics of creation, operation, and destruction of the community as a new “greater
individual” able to exert purposeful active social behaviors—examples of which are the
collective strategies discussed in (Astley & Fombrun, 1983). In (De Florio & Blondia, 2010)
we used the term social energy to refer to the product of such collective behaviors. In the same
paper the term “community” was used to refer to social organizations.
...with channels of communication: Such channels represent the media by means of which
the constituents of social organizations timely share their individual goals, situations, and
states. Channels also induce concepts such as proximity and membership: Depending on the
characteristics of the communication channels members of the communities shall or shall not
be able to access shared knowledge and take part in social decisions.
The just mentioned aspects all play an important role in a specialization of Boulding’s social
organization that we called Fractal Social Organizations (FSO). In order to present the specific
aspects characterizing FSO with respect to its genus proximum in what follows we first introduce,
in Sect. 2.1, a number of preliminary concepts by means of a case study. Secondly, in Sect. 2.2,
we enunciate a number of architectural elements underlying those preliminary concepts. Finally, in

Sect. 2.3 and Sect. 2.4, we make use of the above concepts and elements to define respectively the
building block and the control architecture of our FSO.
2.1. Case Study
Let us suppose that Jane, an elderly woman, is living in her smart house. Jane’s smart house includes
several devices, among which an accelerometer that is used to assess situations such as “Jane has
fallen”. The smart house service includes also a general practitioner (GP) who is timely informed
of situations such as the above one through some communication channel. Let us consider the
following set of roles:
S1 = {δ0, δ1, δ2, . . . , δd, GP, Channel, Jane},
in which δ0 is the above mentioned accelerometer, δ1 is an alarm module (to be introduced later on),
Channel is a communication channel to transfer information between δ1 and GP, and δ2, . . . , δd are
devices that though available are not relevant to the current case. We shall refer to sets of roles as to
societies.
Let us now suppose that situation s = “Jane has fallen” takes place. Let us assume that δ0 operates
according to its specifications and correctly assesses the situation at hand. We shall then say that δ0
becomes active. According to Boulding’s definition of a social organization this implies that S1 is
now partitioned into two running components:
1. The running selection of roles appointed to deal with s (let us refer to this series of selections
as R1
1, R2
1, . . .)
2. The remaining roles (namely, the set of merbers that are inactive with respect to s, identified
by sets L1
1, L2
1, . . .).
In the case at hand the initial partition is given by
L1
1 = S1 {δ0}, R1
1 = {δ0}.
As δ0 is part of the current set R we shall say that δ0 is currently active. Becoming active means
that δ0 starts looking for pertinent activities—activities that is that are relevant to s. Activities are
interpreted here as well-defined protocols to deal with the situation at hand. In what follows we
shall not consider the nature of these protocols and just assume that they are formal descriptions of
common practice, specifications, or regulations, that have been associated to the roles of S1 through
some mechanism (for instance, meta-data).
In what follows we shall assume that one or more “reference domains” (RD) are semantically
associated to activities and societies, and that selections of societies inherit the RD of their superset.
In what follows we shall refer to selections of a society S that are active with respect to a given
situation s as the communities of S with respect to s. S and s will not be mentioned when obvious
from the context.
In the rest of this section we shall denote activities as follows:
¾situation : action¿.
Actions shall take the form:
( role → step ) ,
where “ ” stands for “one or more occurrence”. Occurrences are separated by either “;” (for
sequential execution) or by “//” (for parallel execution). Parentheses may be used to group
occurrences.
We now assume the presence of the following activity among those pertaining to role δ0:
¾fallen : (δ1 → alarm(fallen))¿. (1)
Activity (1) states that, once δ0 “fallen” is detected, δ1 is to execute a single action step and raise
an alarm. It is assumed in what follows that for said activity RD is equal to string “Healthcare”.

It is worth remarking here how action steps call for roles to be played by some actants. If a
corresponding role can be found in community R1
1, that role is associated with the execution of the
corresponding step. On the contrary a restructuring takes place. As an example, if “alarm(fallen)”
calls for a communication channel, then this will lead to the following new selections:
L2
1 = S1 {δ1, Channel}, R2
1 = {δ1, Channel}.
The resulting new community R2
1 can now execute (1). After doing so, R2
1 dissolves back into the
original society: L3
1 = S1, R3
1 = ∅.
We now suppose that activity (1) injects the following new situation: “Alarm has been triggered”.
The corresponding activity is assumed to be the following one:
¾alarm(fallen): ( ( GP → fallenGP(fallen) )
// ( neighbor → fallenNeighbor(fallen) )
// ( relative → fallenRelative(fallen) ) )¿.
The first step of said activity implies updating the L and R selections as follows:
L4
1 = S1 {GP, Jane}, R4
1 = {GP, Jane}.
The treatment of this case is similar to what discussed above and therefore it will not be
repeated. What we now focus our attention on is the fact that action steps such as “(neighbor →
fallenNeighbor(fallen))” can not be resolved within the current society: In fact S1 does not include
a “neighbor” role. When a society finds itself short of a role, a new “meta-situation” occurs and
triggers the following default activity:
¾∅ : (S1 → up(neighbor, alarm(fallen)))¿. (2)
Situation “∅” may be interpreted as an exception that either forwards the request to any superset
societies that include the current one (if at least one such society exists) or it fails (if no superset
societies can be found). It is assumed that said selection is operated by creating a ranked list of
the superset societies and selecting the top “best matching” candidates. Ranking is operated by
considering a metric function measuring a “distance”, i.e., a compatibility degree, between the RD
of the activity and those associated to superset societies. Mission requirements, e.g. geographical
proximity, may also be used to operate said ranking.
Now let us suppose that at least one compatible superset society does exist and be equal to
S2 = {S1, Neighbor, Something else}. (3)
Through the onset of the meta-situation in S1, roles Neighbor and S1 become active. As usual this
translates into a partitioning of S2 into the following two blocks:
L1
2 = {Something else} and R1
2 = {S1, Neighbor}.
As a consequence, Neighbor now can become active within R1
2 and deal with the situation
inherited from S1. We remark how this mechanism correspond to a change of scale. Note also
how the life span of community R1
2 is defined by the duration of the action step “neighbor →
fallenNeighbor(fallen)”.
Intra- and inter-community social activities—we conjecture—may be useful to express autonomic
collaboration services among organizations. A distinctive advantage of such services would be,
e.g., the intelligent and timely resource sharing among care departments and institutions (Waring
et al., 2006). An interesting system based on such principles is SHINE (Secured Health Information
Network and Exchange), a web and mobile-based system for inter-facility health e-referrals†
.
†SHINE enables “health care providers and facilities to efficiently operate and communicate individual or aggregated
referral and case records to the right people in a timely, accurate, and interactive manner” (Anonymous, 2013b). Operating
in the Philippine, SHINE for instance eliminated “the long hours that a pregnant mother spends while looking for a
hospital on a trial-and-error basis” (Valmero, 2011).

Lessons learned from the above simple case study allow us to identify already the following
preliminary list of system requirements to the management of complex collective behaviors:
• Services and policies are required to associate social actants to roles.
• Actions and protocols are required to express and communicate situations, activities, and
actions.
• Services are required to manage the dynamic life-cycle of communities.
• A hierarchical organization of societies and communities as well as services for managing
exceptions and the ensuing propagation of situations are required.
The above requirements are complemented in what follows with a number of architectural
elements and assumptions.
2.2. Architectural Elements and Assumptions
In previous section we introduced the concepts of exceptions and of hierarchies of societies and
we identiﬁed a number of system requirements. Here we go one step further and sketch the basic
elements and assumptions of an architecture based on the principles and requirements discussed so
far. In what follows we shall purposely not distinguish between individual and social entities‡
.
Element 1 (Permanent and transient communities)
Let us consider two classes of communities: Permanent and transient communities. A community
is permanent when it is to respond to a persistent situation; an example is given by hospitals, which
answer the persistent necessity to provide medical care to people in need. Another example is S1 in
Sect. 2.1. On the contrary a community is said to be transient when it is dynamically constructed so
as to respond to the onset of a new situation; subsets R1
1–R4
1 in Sect. 2.1 are all examples of transient
communities.
Element 2 (Permanent and transient roles)
Let us consider two classes of roles: Permanent and transient roles. A permanent role is one
constitutionally assigned to a permanent community—e.g., the logistics department of a hospital
or the hospital itself. A role is said to be transient when it is played by a transient community or
when it is different from those constitutionally assigned to a permanent community.
Element 3 (Hierarchical structure)
Societies and communities shall be structured as the blocks of a Russian nested doll (also known as
“matryoshka doll”), with the following peculiar differences:
• Blocks may belong to more than one doll at the same time. This assumption corresponds to
the hypothesis that actants may play at the same time different roles in different communities.
• Blocks may contain more than one other block.
• Permanent blocks are semantically annotated so as to publish one or more of their RD
(reference context domains) as well as other information. As an example, through this
mechanism a hospital block may publish its nature of “care organization” and for instance
its policies of intervention.
Element 4 (Roles as references)
Blocks shall be represented as multisets of references to roles. Creating a new block from the roles
in an existing block does not subtract the roles from the latter—it only copies their associated
references.
Transient communities dynamically create new temporary blocks (i.e., new active R subsets)
whose lifespan is the amount of time required to deal with their associated situations—e.g. “fallen”
in (1). As mentioned in Sect. 2.1 such processing may result in two cases:
‡As well known, e.g., a human being may be regarded as either an individual entity or a social “body” of specialized
modules— the organ systems. In turn, each organ system may be considered as either an individual—that is, non-
dividable—module or another example of social organizations.

1. The temporary block can be created making use exclusively of roles available within the
parent block. In this case the temporary block becomes a new sub-block of the parent block
until it concludes processing its situation.
2. The temporary block requires roles that are available without the parent block. Through the
exception mechanism discussed in Sect. 2.1 the situation and the call for roles is forwarded
to the parent of the parent block. If the parent block has more than one parent, a selection
is carried out by considering the reference context domains as well as possible constraints at
individual, collective, and mission level. At the current, still abstract level of specification, we
shall not discuss in detail how the selection is carried out and how the missing role is actually
sought.
Once all required roles are found, a new block is constructed.
Element 5 (Role templates)
Some mechanism may be foreseen such that, though transient in nature, temporary blocks leave
a “trace” in the system—thus becoming new known “templates of roles”. If a certain complex
situation regularly repeats itself then a corresponding role template may be recalled and considered
as a more compact (and thus, efficient) way to deal with it.
Role templates result in an elementary form of learning through experience, which corresponds to
the ability to respond more rapidly and efficiently to challenging environmental situations. Instinct
(innate behaviour) may also be based on predefined and genetically encoded role templates. We also
conjecture that the adoption of mechanisms similar to our templates of roles may be associated to
the widespread emergence of modularity in nature (Clune et al., 2013).
More complex learning solutions may involve “scoring” the templates of roles with respect
to their effectiveness while servicing a given recurring situation. By choosing the currently best
matching candidate and by updating the templates’ scores systems based on such strategy may
autonomously improve their environment fit. A similar scoring mechanism may be applied when
appointing the actants to be associated to sought roles—as it was done, e.g. in (Buys et al.,
2011; Buys et al., 2012), to self-select the best software versions for an N-version programming
composite (Avi˘zienis, 1995).
Element 6 (Template roles)
A template of roles shall be considered as a special case of role. Similarly to the concept of
subroutines in programming languages, template of roles may provide a way to refer concisely
to complex roles and their corresponding activities.
Through the above assumption templates of roles allow complexity to build up into more and
more “high level” roles (templates of templates, and so forth), which correspond to more and more
abstract activities. Intuitively, such activities may be interpreted as the tokens of the higher level
“languages” used by complex communities to play their roles.
Element 7 (Role permanentification)
When a template of roles (or of other templates) is repeatedly selected as an answer to a recurring
situation, the corresponding role shall become permanent. A corresponding set of actants that
specializes to play the template role may then be labeled as a new permanent community.
We now go one step further and make use of a subset of the just discussed requirements and
architectural elements to sketch the elements of the societal “building block” of our FSO as well as
its hierarchical organization.
2.3. Service-oriented Communities
Service-oriented Community (SoC) is the name we use to refer to the architectural element
corresponding to the concepts of “community” and “block” introduced in previous sections. As
described in (De Florio & Blondia, 2010), a SoC is a collective socio-technical system coupling
services provided by smart cyber-physical “things” with services supplied by human beings. Actants
that agree to join a SoC become their “members”. Members of a SoC are diverse, which translates
into a rich variety of services. By focusing for instance on human members only, diversity implies

• different know-hows (e.g. those of a general practitioner, or those of a gardener);
• different policies for providing services (e.g. well-defined time schedules and fares, or
dynamically varying availability to provide free-of-charge services as occasional informal
carers);
• different location, in that members may be mobile, thus able to get dynamically closer to or
farther from other members;
• different value systems,
and so on. Such attributes are called in what follows features. Finally, members may have different
goals—for instance being able to reach a given location within a certain amount of time with a
certain budget of travelling costs and with at least a given quality of experience.
We shall call viewpoints any well-defined and agreed representation of a member’s features and
goals.
The operational model of the SoC is represented in Fig. 1, in which members publish their
viewpoints to other members in logical or physical proximity. Viewpoints may be issued, e.g., by
an elderly member in need of assistance—such as Jane in Sect. 2.1,—by a bystander posting their
tweets to describe an accident, by a monitoring device notifying the occurrence of some critical
event, by the acceptance service of a hospital, and so forth. Control and adaptation are achieved by
sharing viewpoints, unravelling semantic analogies among them and creating transient “networks
of relations” (Latour, 1996) among members. A way to enact the above semantic processes is
described, e.g., in (Sun et al., 2007).
Following some strategy (e.g., the one described in (De Florio et al., 2000)) one or more members
are given the responsibility to host a viewpoint registry and act as temporary coordinator. The
coordinators then become “collective members,” namely members representing a whole community
of other societal actants. Among their duties, the coordinators are to identify whether the viewpoints
currently stored in the registry correspond to a known situation such as those discussed in Sect. 2.1.
Analyses are performed by semantically matching new viewpoints with those already stored in the
registry (De Florio & Blondia, 2010). Analogies between viewpoints are identified as shown in (Sun
et al., 2007) by making use of ontologies associated with the reference domains. Once a similarity
is discovered the coordinator notifies the corresponding members. The shared knowledge enables
the creation of “bindings”, which may be spontaneous or mandatory depending on the context and
the policies in use. A practical way to implement said bindings in a flat (that is, non-hierarchical)
community was the so-called “Participant mode” described in (Sun et al., 2007).
Through bindings, members create a new temporary community. In conformity with what
assumed in Sect. 2.1 and Sect. 2.2 the coordinators also register the association between members
and roles and appoint the execution of activities. Coordinators may also keep track of the
effectiveness of the enacted communities in dealing with experienced situations through the “role
templates” of Sect. 2.2, which could be labeled with a running “score” tag. As already suggested, the
same principle could be applied in the selection of the actants to be enrolled for a new community
derived from a role template.
An example of a Service-oriented Community specifically targeting Ambient Assistance Living
services is given by the Mutual Assistance Community (Sun et al., 2010; Sun et al., 2007; Gui et al.,
2007). We refer the reader to the just cited references for more information on the technical design
of a SoC.
2.4. Fractal Social Organizations
A FSO may be concisely described as a fractal organization whose building block, or canon, is a
Service-oriented Community. The rest of this section in fact may be considered as an explanation
of the just given definition. More information about fractal organizations and the concept of canon
may be found in Sect. 4.1.
Fractal organizations are bio-inspired hierarchical distributed control architectures whose
components are simultaneously an individual and a social entity, “a part and a whole, a container and
a contained, a controller and a controlled” (Sousa et al., 2000). Said components are autonomous
entities corresponding to the communities exemplified in Sect. 2.1 and Sect. 2.2 and implemented

Figure 1. Representation of the canon of the Service-oriented Community.
as service-oriented communities as sketched in Sect. 2.3. By allowing SoC to become members
of other SoC we come up with an architecture with the “matryoshka doll” structure suggested in
Sect. 2.2, Element 3. Through the coordinator role the SoC assume a twin nature: They are at the
same time social collectives and social individualities, as exemplified by S1 in (3). This process is
known as “personization” in Actor-Network Theory (Latour, 1996).
The coordinators are also responsible for declaring the onset of exceptions (see (2)) and for
propagating the corresponding unresolved situations. This means forwarding the event to all
superset communities the current community is a member of, and possibly to other nearest
neighboring “higher-level” SoC. The “reference domain” meta-data may be used to define a metric
function as suggested in Sect. 2.1. This is represented in Fig. 2.
An important aspect of adopting SoC as building block for our fractal social organizations is that,
coherently with what discussed in Sect. 2.1, it allows multiple redundant responses to be considered
and arranged at the same time. As an example, a request from an elderly man in need of assistance
may be answered by (a) professional practitioners, (b) informal carers, (c) his relatives, (d) his
neighbours, and so forth. These multiple, redundant “planes” of assistance supply independent
service routes, which may be complemented by traditional (consolidated) service providers—e.g.
social insurance companies or other healthcare bodies. Furthermore, this approach makes it possible
to select a Pareto-optimal “plane” of assistance corresponding to the solution which provides the
expected quality of services with minimal resource consumption. Evidence to this second statement
has been obtained by simulating the social interactions of a single, non-hierarchical community (Sun
et al., 2007). In the same reference we showed how unraveling analogies via semantic reasoning
promotes both a higher level of resource utilisation and a stronger quality of experience.
An exemplary FSO is depicted in Fig. 2, which portrays service-oriented communities at different
scales—here identified as Layer 0, Layer 1, and Layer 2 members. A second representation is shown
in Fig. 3. In this case the FSO layers include individuals, smart houses, hospitals, and inter-facilities
of care. The same triangle-shaped pattern of the SoC repeats itself at each scale. The right-hand side
of Fig. 2 provides a representation of this phenomenon in terms of a well-known fractal structure
whose dimension is approximately equal to 1.5849625.

Figure 2. The top picture represents the SoC architecture. Resemblance with a known fractal is shown in the
bottom picture.
We have just sketched the main characteristics of FSO, a socio-technical system with a complex
fractal organization. As a preliminary way to evaluate the effectiveness of our architectural
choices we now introduce a formal model based on Boulding’s definition of social organization
(see Sect. 2). A remarkable ensuing result is that several of the architectural elements of our
Fractal Social Organization—including, e.g., self-similarity, the appearance of modular structures,
and a hierarchical structure resulting from the recursive application of building blocks—emerge
spontaneously.
3. ALGEBRAIC MODEL FOR SOCIAL ORGANIZATIONS
As we mentioned in Sect. 2, Fractal Social Organizations may be considered as a specialization
of Boulding’s social organizations, which in turn are defined as “a set of roles tied together with
channels of communication”. Inspired by the algebraic nature of Boulding’s definition and by
previous work on the dynamics of multisets (De Florio, 1995; De Florio, 2005) we introduce in
what follows a formal model for the dynamics of the behaviors originating in social organizations.
Said model is presented in Sect. 3.1 while its properties are discussed in Sect. 3.2.

Figure 3. Exemplification of SoC social organizations.
3.1. Formal Model
Let us assume we have a set of roles, uniquely identified by integer numbers, as well as a set of
actants, each of which is associated with exactly one of the above roles. In what follows for the sake
of simplicity of treatise we shall assume this association to be static. As an example, let us say we
have
2 GPs, 2 nurses, and 8 patients,
collectively identified by multiset
S = {0, 0, 1, 1, 2, 2, 2, 2, 2, 2, 2, 2}. (4)
In accordance with what discussed in Sect. 2.1 we shall refer to multiset (4) as to a society.
Furthermore, let us consider a concatenation operator, ·, to build ordered sequences of roles from
a society. In what follows we shall refer to any such ordered sequence as to an organization. An
example of organization is the following sequence:
0 · 1 · 1 · 2 · 2 · 2 · 2 · 0 · 2 · 2 · 2 · 2.
We shall also refer to the above sequence as “0 · 12
· 24
· 0 · 24
”, or simply as “011222202222”.
No specific meaning is given in what follows to organizations, their purpose, or their behavior.
The only aspect that is highlighted in the above definition is the order and role of its constituents,
as it is the case e.g. in DNA sequences of nucleotides. Another exemplification is an assembly line
or a construction pipeline, where “something is done” by moving objects through a string of actants
identified by their role (i.e., their peculiar function).
Now, as we did in Sect. 2.1, we shall consider the onset of some situation (Ye et al., 2012)
(endogenous or exogenous to S) such that some reaction is triggered by or through some of the
actants in that society. Said situations may be interpreted as opportunities (e.g., the onset of some
competitive economic advantage), threats (e.g., the outbreak of an infectious disease), crises (as a
result, e.g., of a stock market crash), or other circumstances or changes.
Consistently with what we have discussed in Sect. 2.1 here we assume the onset of s, namely
situation “one of the patients has fallen”. Through some communication channel we further assume

that situation s triggers the intervention of 1 GP and 1 nurse. As a consequence of said intervention,
society (4) gets now partitioned into the following two blocks:
subset L = {0, 1, 2, 2, 2, 2, 2, 2, 2} and subset R = {0, 1, 2}, (5)
in which block L is inactive with respect to s, while R is active with respect to s. In what follows we
shall use the same terminology we introduced in Sect. 2.1 and refer to an active subset of a society
as to a community. When obvious from the context the firing situation will be omitted. Being active
means that the actants in R shall need to organize themselves (in some sense and here not specified
way) so as to deal with the situation at hand. Summarizing, multiset S is partitioned with respect
to s into an inert subset L and some organization of community R—that is, some ordering of the
active actants in S.
In what follows we shall model the dynamics of multisets L and organizations R. We shall assume
that events occur at discrete time steps and are modeled as either processing events or perturbations:
The former case stands for functional events corresponding to the processing of some workflow
internal to the organizations, the latter is the onset of new situations leading to
1. either a reorganization of the elements within a community
2. or a repartitioning of the society into two new blocks L and R .
In what follows we model the reorganization in step 1 as a permutation of R—that is, a new
ordering for its constituents. The rationale of this is that in this case the community responds to the
onset of s by finding resources within itself, possibly reshaping the flow of activities, but keeping
the same roles. In mathematical terms we say that R is closed with respect to s.
The repartitioning in step 2 means that the community calls for external resources—resources
that were inactive (i.e., they correspond to roles in L) but now need to be enacted (entering R and
thus constituting a new R ).
Let us assume that a given society consists of r roles, identified by numbers 0, . . . , r − 1, and that
role i is played by ni actants, 0 ≤ i < r. Let us call “First” a function defined as follows:
First : P(S) → O(P(S)), such that
∀X ⊂ S : ∃n0, n1, . . . , nr−1 : First(X) = 0n0
· 1n1
· (r − 1)nr−1
, (6)
where P(S) is the powerset of S (i.e., the set of all the subsets of S) and O maps sets onto
organizations. Note that First generates the organization corresponding to the “smallest” number
whose digits are the role identifiers. Let us refer to First(X) as to the “first organization” of X.
Similarly we define dual function “Last” and refer to Last(X) as to the “last organization” of X.
We now recall the definition of function “Succ” (adapted from (De Florio, 2005)). Succ takes as
input sequence
First(L) · R (7)
and returns the sequence corresponding to the next organization in the lexicographically ordered set
of all organizations of the input sequence. Applying Succ to the last organization returns the first
one. Dually, we define Succ−1
as the function returning the previous organization. If applied to the
first organization, Succ−1
returns the last one.
In principle any function producing all the organizations from a reference society could have been
used instead of Succ, but making use of the lexicographic order has the advantage that it makes it
easy to produce numerical representations that are all positive and monotonically increasing. Some
of the geometrical representations introduced in Sect. 3.2 require the latter property.
We now can model both cases of perturbations by “tossing a coin” corresponding to a non-zero
relative integer number, say z, and applying (Succ)z
if z > 0 and (Succ−1
)−z
if z < 0§
.
§In fact, due to the “wrap around” of functions Succ and Succ−1, z may be substituted with z mod pS, where pS is the
number of permutations of multiset S. This number also represent the amount of distinct orbits of Succ and Succ−1, i.e.
the length of their only cycle.

Figure 4. Exemplary life-cycle for the organizations of S = {0, 0, 1, 2, 3}.
.
If we toss a coin and apply the above process at each time step t, the following series
L(t), R(t) t≥0
(8)
defines a dynamic system corresponding to the dynamic life-cycle of all the possible organizations
originating from a given reference society of roles. Figure 4 shows all the orbits of (8) for society
{0, 0, 1, 2, 3} as well an exemplary life-cycle of a community from that society.
It is also possible to define RS as a relation between the organizations from a given society S
such that, for any two organizations of S, say a and b,
a RS b iff Succ(a) = b. (9)
When doing so, the series of orbits in Fig. 4 represent the transitive closure of RS. Note also how (8)
may be described as a “random walk” through that series. Figure 4 depicts one such exemplary
random walk.
3.2. Representations and Properties
The just described model considerably extends the deterministic combinatorial model introduced
in (De Florio, 2005). As it was done in the cited paper, here we can provide geometrical
representations for the evolution of the dynamic system in (8). In what follows we introduce some
of said representations.
• A first representation is obtained by mapping successive orbits of Succ onto integer numbers.
This is done by interpreting roles as digits and using the number of available roles, r, as base.
We call the obtained integers “organization numbers”. We shall use ν to represent the just
introduced mapping. Figure 5 shows this representation for society S = {03
, 112
}.
• A second geometrical representation is derived by taking the difference between two
consecutive organization numbers:
∀o ∈ O(S) : δ(o) = ν(Succ(o)) − ν(o).
We refer to the obtained integers as to “delta steps”. Figure 6 shows the delta steps of
society (4).
• Figure 7 shows the logarithms of delta steps.

Figure 5. Organization numbers of society S = {03
, 112
}. Abscissa correspond to the successive orbits of
Succ (no wrap around is shown). Ordinates are the corresponding organization numbers. Note how dynamics
such as these are self-similar in that they include the dynamics of smaller societies.
• Finally, a family of geometrical representations is obtained by breaking down organizations
into m consecutive “chunks” of roles and by interpreting each of them as a base-r integer. In
turn, these numbers are used to identify points in m-dimensional space. Figure 8 shows this
representation for m = 2 and society (4) while two examples for m = 3 are depicted in Fig. 9
and Fig. 10.
As evident from the just referenced figures, several of the above representations are characterized
by some degree of self-similarity and by the emergence of modularity. This is particularly
meaningful in view of the results of other researchers. Johns, for instance, remarked in (Johns,
2011) how self-similarity is a necessary condition to self-organization, while the widespread degree
of modularity in the natural world has been associated by many with the emergence of evolvability—
the ability to rapidly adapt to a turbulent environment (Clune et al., 2013).
A second property of some of the just introduced geometrical representations is the fact that they
are factorisable: Their structure can be shown to be governed by well defined rules reproducing
the dynamics of simpler and simpler organizations, down to some atomic or “prime” organizations
that cannot be further simplified. This is shown for instance in Fig. 11. An interpretation of this
phenomenon is that factorization represents a change of scale: Prime organizations unite into a
coherent and self-similar hierarchical organization. The higher we go in the scale the more complex
is the organization. Such complexity though is not introduced arbitrarily but according to a well-
defined general rule. Remarkably, the above behaviors and the resulting “structured addition” of
complexity characterizing factorisable organizations are at the core of the concepts of Fractal Social
Organization and Service-oriented Community introduced in previous sections. It is our conjecture
that the above emerging traits and properties may hint at the emergence of scalability and robustness
in future socio-technical complex systems designed after our Fractal Social Organizations.
A third noteworthy property of some of our geometrical representations is revealed when
estimating their fractal dimension. For this we found a useful tool in Fractalyse (Vuidel, 2013), a
software package that provides several classical methods to measure the fractal dimension of black-
and-white images such as the one shown in Fig. 8. Available methods include box counting, radius
mass, correlation, dilation, and others. Fractalyse was successfully applied to estimate, e.g., the
fractal dimension of urban areas in Europe and China (Ma et al., 2008). Figures 12–15 show some

0
5000
10000
15000
20000
25000
30000
35000
0 500 1000 1500 2000 2500 3000
DeltastepsofS={0,0,1,1,2,2,2,2,2,2,2,2}
Figure 6. Delta steps of society S = {02
, 12
, 28
}. Also in this case self-similarity can be observed.
Figure 7. Logarithms of delta steps for society {0, 1, 2, 3, 4, 5}. Self-similarity is highlighted by emphasizing
the representations corresponding to society {0, 1, 2, 3} (small-sized region) and society {0, 1, 2, 3, 4}
(medium-sized region).
of the results we obtained with Fractalyse. As an example, the Box counting method with parameters
y = a × xd
+ c and exponential box size set to 2, applied to the bidimensional representations
of society S = {0, 1, 2, 3, 4, 5, 6, 7}, estimates with nearly perfect correlation a fractal dimension
d ≈ 1.792 with a = 9.5377 × 10−3
and c = 4.5496 (Fig. 14).
4. RELATED CONCEPTS
Three major concepts related to FSO are the focus of this section, namely (1) organizational aspects,
and in particular their nature being bio-inspired distributed organizations; (2) knowledge-related
aspects, namely their relation with so-called knowledge ecosystems; (3) the FSO social dynamics,

Figure 8. Bi-dimensional representation of the organizations in S = {04
, 12
, 24
} (equal sized chunks).
whose closest inspiration comes from the social theory known as Actor-Network Theory (Latour,
1996). In what follows we highlight very concisely the relationships between FSO and the above
mentioned disciplines and subjects.
4.1. Organizations
A fundamental aspect for the effective emergence of desirable properties and behaviors—e.g.
scalability, manageability, robustness, and resilience—is “the way a social union is structured and
organized” (De Florio, 2013). Common terms to refer to such concept are “Partnership”, “reference
architecture”, “organizational structure”, “control structure”, and simply “organization”—the latter
being the term we shall use in what follows.
Commonly employed organizations include:
• Centralized control organizations, characterized by a single point-of-control-and-maintenance
(PCM). The PCM is the “personization” of the whole organization. Simplicity is the main
advantage of centralization, which is paid back by the introduction of a single-point-of-failure
as well as a structural bottleneck.

Figure 9. POV-ray (Froehlich, 2012) rendition of society (4) (equal sized chunks).
Figure 10. POV-ray picture of S = {0, 1, 27
} (equal sized chunks). Note how the depicted structure is one
of the recurring patterns in Fig. 9. Note also how this closely corresponds to the concept of role template
introduced in Sect. 2.2. This is because S is in fact a subset of society (4) (as shown e.g. in (5)).

Figure 11. Society {08
, 16
}, when represented bi-dimensionally, can be factorized into several instances of
atomic societal nuclei, or “prime communities”. Note how this closely correspond to the concept of canonical
“building blocks” such as the Service-oriented Community introduced in Sect. 2.3. (Picture from (De Florio,
1995)—copyright by Wolfram Research Inc. Used by permission.)
Figure 12. The Correlation method of Fractalyse applied to the bidimensional representation of society S =
{0, 1, 2, 3, 4, 5}. The analysis reveals a fractal dimension of 1.704 with a correlation coefficient c ≈ 0.9999.
• Hierarchical control organizations are characterized by a top-down flow of commands
and a bottom-up flow of feedback information. These flows are limited to the immediate
lower/upper level. Hierarchies scale better than centralized organizations though are not free
of shortcomings; in particular information must flow throughout the hierarchy to reach the top
control level, which translates in propagation delays and possibly propagation failures (viz.
loss or corruption of commands and feedbacks).
• Heterarchical control organizations are those in which autonomous agents coexist in a flat
structure with no predefined relationships. Information is distributed and there is no global
knowledge nor fixed points of control. Such “heterarchies” (Stark, 1999) are characterized
by the emergence of adaptability, by structural adoption of diversity, as well as by the
ability to overcome so-called “lock-ins” (local minima of the system-environment fit).

Figure 13. The Radius Mass method of Fractalyse applied to the bidimensional representation of society
S = {0, 1, 1, 2, 2, 3, 3, 4}. A fractal dimension of 1.506 is suggested with a correlation coefficient c ≈ 0.993.
Figure 14. The Box Counting method of Fractalyse applied to society S = {0, 1, 2, 3, 4, 5, 6, 7}. Results
suggest a fractal dimension of 1.792 with a correlation coefficient c ≈ 0.9999.
A major shortcoming of heterarchies is that “central scheduling or resource planning is
impossible” (Ryu, 2003).
Other organizations emerged as an attempt to reduce the shortcomings of the above mentioned
ones. The main characteristics of said organizations are their bio-inspired origin and their distributed
nature. Such class of mechanisms is particularly interesting in the context of this paper as it has
been successfully applied to design of systems exhibiting resilience and robustness in the face of
changes and failures. In particular here we shall briefly recall the characteristics of three classes
of distributed control mechanisms—bionic, holonic, and fractal architectures (Tharumarajah et al.,
1996; Ryu, 2003).
Bionic organizations are based on a hierarchical composition of autonomous building blocks
characterized by spontaneous behavior and local interaction called cells or modelons. Biologically

Figure 15. Scaling behavior and error curve produced by Fractalyse with the method of Fig. 14.
inspired concepts such as enzymes and hormones are used to model conflict resolution schemes.
Attraction and repulsion fields are used to model collaborative tasks.
Holonic organizations are compositions of building blocks, called holons, which “are
simultaneously a part and a whole, a container and a contained, a controller and a controlled” (Ryu,
2003). A same structure and a same set of configuration rules (the already mentioned called “canon”)
is repeated at different granularity scales producing so called holarchies. Holons are autonomous
entities that establish cooperative relationships and are characterized by the emergence of stability,
flexibility, and efficient use of the available resources. An example of said holarchies is given by
so-called holonic manufacturing systems (Ryu, 2003).
Fractal organizations are similar to holarchies, but the canon is not statically defined in that it may
freely evolve and differentiate. Local interaction and experience produce custom restructuring and
regrouping as exemplified, e.g., in the Fractal Company (Warnecke & Hüser, 1993) and the Fractal
Factory (Tharumarajah et al., 1998).
4.2. Knowledge Ecosystems
Knowledge ecosystems (KE) are social organizations that aim at the emergence of collective
intelligence through the mutual promotion of the individual and the social dimension. KE are based
on the principle of social constructivism—the hypothesis that learning is a social process that takes
place through the interactions in a society. In order to consolidate knowledge and produce wisdom,
it is key that “information, ideas, insights, and inspiration cross-fertilize and feed one another
free from the constraints of geography and schedule¶
”. KE comprise three overlapping layers
or networks: (1) A social network whose members’ “productive conversations” and collaborative
behaviors create (2) “a knowledge network of ideas, information, and inspiration”, sustained by (3)
a technology network to communicate and persist ideas, knowledge, and lessons learned through
experience. In order to function, KE must enable the quick sharing of the above ideas, information,
and inspiration owned by the participating people, which is done through the technology network.
This process is called by Pòr “electrification” of the knowledge network. Through such process
the social network becomes a “web of distributed intelligence”—a sort of collective “nervous
system” that is to enhance an organization’s ability to tap into its own collective intelligence and
thus formulate the most appropriate responses to turbulent environmental conditions. Said nervous
¶Here and in the rest of Sect. 4.2 quotes are from (Pór, 2000).

system is “embedded not in computers and hardware, but in the interactions among people that bring
the organization into existence day after day”.
4.3. Actor-Network Theory
Actor-Network Theory (ANT) is a complex social theory based on the above mentioned hypothesis
of social constructivism and the central idea that “essences” (viz., individuals and societies) are to
be interpreted not as “containers” characterized by a physical dimension, e.g., a surface or a sphere,
but rather as networks of nodes that have as many dimensions as they have “ties” (i.e., connections) .
Such ties are “weak by themselves”, though they achieve robustness (“material resistance”) through
their social nature: “Each tie, no matter how strong, is itself woven out of still weaker threads [..]
Strength does not come from concentration, purity and unity, but from dissemination, heterogeneity
and the careful plaiting of weak ties”. “Strength” here refers to the ability of the “essences” to
retain their identity in spite of environmental conditions affecting their ties and nodes. A fragile
essence is one characterized by one or more points-of-diffusion-failures—as it is the case for the
centralized and the hierarchical organizations discussed in Sect. 4.1; conversely, an essence is robust
if it tolerates discontinuities and other information diffusion failures.
Be it an individual or a society, an ANT essence is not a static, immutable entity: It “starts
from irreducible, incommensurable, unconnected localities, which [..] sometimes end [up] into
provisionally commensurable connections”. Strength is sought by conserving identity despite the
changes of scale that are necessary to counterbalance turbulent environmental conditions. The above
mentioned “careful plaiting of weak ties” is meant to guarantee that a network “is the same”, though
“stronger”. The structured addition of complexity that we observed in the organizations discussed in
Sect. 3 may provide—we conjecture—a geometrical interpretation to the ANT concept of strength.
5. CONCLUSIONS AND NEXT STEPS
We sketched the main design elements of fractal social organizations—a class of socio-technical
complex systems structured as fractal organizations and based on the recursive application of a
same building block, the service-oriented community. Self-similar, modular, and fractal-like “by
construction”, our solution—we argue—may provide our societies with a scalable and robust
system structure able to match effectively the dynamically varying requirements and the turbulent
environmental conditions that more and more threaten the manageability of traditional human
organizations (Barabasi et al., 2013).
As a preliminary way to prove the meaningfulness of our design choices we introduced a formal
model for the dynamics of simple, non-hierarchical social structures organized as permutations of
multisets of roles. Remarkably enough, we observed how the simple assumptions in our model
resulted in the spontaneous emergence of a hierarchical and modular organization characterized
by a structured addition of complexity and a fractal nature closely resembling the design traits of
our organizations. By virtue of said resemblance and building on top of past results (Johns, 2011;
Clune et al., 2013) we conjecture that the concepts presented in this paper may be used to capture
and conquer the complexity of traditional organizations that address vital services of our societies
including, e.g., care, crisis management, goods and energy distribution, and civil protection. In
the long run, should our conjecture prove correct, this may help fulﬁlling the visions of “smarter
organizations” such as those expressed in (Anonymous, 2013a).
Considerable new challenges will need to be addressed in our future work. This will include
investigating the effectiveness of FSO as a system structure for the intelligent sharing of resources,
competences, knowledge, and goods. Moreover, scientiﬁc evidence shall be sought to verify the
robustness, resilience, and self-management capability of socio-technical complex systems based
on our models. We are already working in this direction by making use of Alchemist (Pianini et al.,
Here and in the rest of Sect. 4.3 quotes are from (Latour, 1996).

2013)—a powerful tool for the simulation and verification of pervasive service ecosystems. Formal
methods such as bi-graphical reactive systems (Milner, 2008) are also being used (Coronato et al.,
2012) to describe the behaviors of fractal social organizations.
Another “grand challenge” we plan to tackle is the design of knowledge ecosystems based
on our models. Here the major property we shall investigate will be the emergence of forms
of collective intelligence as well as advanced collective behaviors including, e.g., collaboration,
collective strategies (Astley & Fombrun, 1983), co-opetition (Brandenburger & Nalebuff, 1998),
and co-evolution (Adner & Kapoor, 2010).
Last but not the least, the current work provides us with a novel interpretation of the phenomenon
of modularization as a spontaneous process governed by simple mathematical rules and initial
assumptions. Further investigation shall be required to verify the relation between our results—and
our “systems based on numbers” (Wolfram, 2002)—and the emergence of evolvability in natural
organisms and biologically inspired artificial organizations (Clune et al., 2013).
ACKNOWLEDGEMENTS
This work was partially supported by iMinds—the Interdisciplinary institute for Technology, a research
institute funded by the Flemish Government—as well as by the Flemish Government Agency for Innovation
by Science and Technology (IWT).
REFERENCES
Adner, R., & Kapoor, R. 2010. Value creation in innovation ecosystems: How the structure of technological
interdependence affects firm performance in new technology generations. Strategic management
journal, 31 : 306–333.
Anonymous. 2011. The 2012 ageing report: Underlying assumptions and projection methodologies.
European economy, 4(September) : 309.
Anonymous. 2012 (October). Population structure and ageing. Tech. rept. Eurostat.
Anonymous. 2013 (a). Smarter planet. Tech. rept. IBM.
Anonymous. 2013 (b). SHINE — Secured Health Information Network and Exchange. Accessed on May 2,
2013 from https://www.shine.ph.
Astley, W.G., & Fombrun, C. J. 1983. Collective strategy: Social ecology of organizational environments.
The academy of management review, 8(October) : 576–587.
Avi˘zienis, Algirdas. 1995. The methodology of N-version programming. Chap. 2, pages 23–46 of: Lyu,
Michael (ed), Software fault tolerance. John Wiley & Sons, New York.
Barabasi, Albert-Laszlo et al. 2013. Perspectives on a hyperconnected world — insights from the science of
complexity. Tech. rept. World Economic Forum’s Global Agenda Council on Complex Systems.
Boulding, Kenneth. 1956. General systems theory—the skeleton of science. Management science, 2(3) :
197–208.
Brandenburger, A., & Nalebuff, B. 1998. Co-opetition — a revolutionary mindset that combines competition
and cooperation.
Bray, David A., Croxson, Karen, Dutton, William H., & Konsynski, Benn. 2008 (February). Sermo: A
community-based, knowledge ecosystem. In: Oll distributed problem-solving networks conference.
Buys, Jonas, De Florio, Vincenzo, & Blondia, Chris. 2011. Towards context-aware adaptive fault tolerance in
SOA applications. In Proceedings of the 5th ACM international conference on distributed event-based
systems (DEBS); 63–74. Association for Computing Machinery, Inc. (ACM).
Buys, Jonas, De Florio, Vincenzo, & Blondia, Chris. 2012. Towards parsimonious resource allocation in
context-aware n-version programming. In Proceedings of the 7th IET system safety conference. The
Institute of Engineering and Technology.

Clune, Jeff, Mouret, Jean-Baptiste, & Lipson, Hod. March 22, 2013. The evolutionary origins of modularity.
In Proceedings of the Royal Society b: Biological sciences, 280(1755).
Coronato, Antonio, De Florio, Vincenzo, Bakhouya, Mohamed, & Di Marzo Serugendo, Giovanna.
2012. Formal modeling of socio-technical collective adaptive systems. In Proceedings of the
first international workshop on adaptive service ecosystems: Nature and socially inspired solutions
(ASENSIS 2012). IEEE.
De Florio, Vincenzo. On the constituent attributes of software and organisational resilience. Interdisciplinary
Science Reviews, 38(2). Maney Publishing.
De Florio, Vincenzo. 1995. The HeartQuake dynamic system. Complex systems, 9(2) : 91–114.
De Florio, Vincenzo & Blondia, Chris. 2008. On the requirements of new software development.
International Journal of Business Intelligence and Data Mining, 3(3) : 330–349.
De Florio, Vincenzo. 2005. Permutation numbers. Complex systems, 15(2).
De Florio, Vincenzo, & Blondia, Chris. 2010. Service-oriented communities: Visions and contributions
towards social organizations. Pages 319–328 of: Meersman, Robert, Dillon, Tharam, & Herrero, Pilar
(eds), On the move to meaningful internet systems: Otm 2010 workshops. Lecture Notes in Computer
Science, vol. 6428. Springer Berlin / Heidelberg.
De Florio, Vincenzo, Deconinck, Geert, & Lauwereins, Rudy. 2000. An algorithm for tolerating crash
failures in distributed systems. In Proc. of the 7th annual IEEE international conference and workshop
on the engineering of computer based systems (ECBS) ; 9–17. Edinburgh, Scotland: IEEE Comp. Soc.
Press.
De Florio, Vincenzo, Coronato, Antonio, Bakhouya, Mohamed, & Di Marzo Serugendo, Giovanna.
2012. Service-oriented communities: Models and concepts towards fractal social organizations. In
Proceedings of the 8th international conference on signal, image technology and internet based systems
(SITIS 2012). IEEE.
DeHon, André, Giavitto, Jean-Louis, & Gruau, Fréderic. 2007. 06361 executive report – computing
media languages for space-oriented computation. In: DeHon, André, Giavitto, Jean-Louis, & Gruau,
Frédric (eds), Computing media and languages for space-oriented computation. Dagstuhl Seminar
Proceedings, no. 06361. Dagstuhl, Germany: Internationales Begegnungs- und Forschungszentrum für
Informatik (IBFI), Schloss Dagstuhl, Germany.
Dilts, D.M., Boyd, N.P., & Whorms, H.H. 1991. The evolution of control architectures for automated
manufacturing systems. Journal of manufacturing systems, 10(1) : 79 – 93.
Dolwick, Jim. 2009. ‘The social’ and beyond: Introducing actor-network theory. Journal of maritime
archaeology, 4 : 21–49.
Fleisch, Elgar. 2010. What is the internet of things? an economic perspective. Tech. rept. WP-BIZAPP-053.
Auto-ID Labs.
Froehlich, Thorsten. 2012. Persistence of vision raytracer on-line documentation repository. Available at
URL http://wiki.povray.org/content/Documentation:Contents.
Gui, N., Sun, H., De Florio, V., & Blondia, C. 2007. A Service-oriented Infrastructure Approach for Mutual
Assistance Communities. In Proc. of the 1st IEEE WoWMoM Workshop on Adaptive and DependAble
Mission- and bUsiness-critical mobile Systems (ADAMUS 2007) IEEE.
Hardin, Garrett. 1968. The tragedy of the commons. Science, 162(3859) : 1243–1248.
Johns, Richard. 2011. Self-organisation in dynamical systems: a limiting result. Synthese, 181 : 255–275.
Latour, Bruno. 1996. On actor-network theory. a few clarifications plus more than a few complications.
Soziale welt, 47 : 369–381.
Ma, Ronghua, Gu, Chaolin, Pu, Yingxia, & Ma, Xiaodong. 2008. Mining the urban sprawl pattern: A case
study on Sunan, China. Sensors, 8(10) : 6371–6395.
Milner, Robin. 2008. Bigraphs and their algebra. Electron. notes theor. comput. sci., 209(Apr.) : 5–19.
Pianini, Danilo, Montagna, Sara, & Viroli, Mirko. 2013. Chemical-oriented simulation of computational
systems with alchemist. Journal of simulation, January. Advance online publication.

Pór, George. 2000. Nurturing systemic wisdom through knowledge ecology. The systems thinker, 11(8) :
1–5.
Rosenblueth, Arturo, Wiener, Norbert, & Bigelow, Julian. 1943. Behavior, purpose and teleology.
Philosophy of science, 10(1) : 18–24.
Ryu, Kwangyeol. 2003 (August). Fractal-based reference model for self-reconfigurable manufacturing
systems. Ph.D. thesis, Pohang University of Science and Technology, Korea.
Sousa, P., Silva, N., Heikkila, T., Kallingbaum, M., & Valcknears, P. 2000. Aspects of co-operation in
distributed manufacturing systems. Studies in informatics and control journal, 9(2) : 89–110.
Stark, David C. 1999. The biology of business: Decoding the natural laws of enterprise (j. h. clippinger
iii, ed.). Jossey-Bass. Chap. Heterarchy: Distributing Authorithy and Organizing Diversity, pages
153–179.
Sun, Hong, De Florio, Vincenzo, & Blondia, Chris. 2006. A design tool to reason about ambient assisted
living systems. In: Proceedings of the international conference on intelligent systems design and
applications (ISDA’06). Jinan, Shandong, P. R. China: IEEE Computer Society.
Sun, Hong, De Florio, Vincenzo, Gui, Ning, & Blondia, Chris. 2007. Participant: A new concept for
optimally assisting the elder people. In: Proceedings of the 20th ieee international symposium on
computer-based medical systems (CBMS-2007). Maribor, Slovenia: IEEE Computer Society.
Sun, Hong, De Florio, Vincenzo, Gui, Ning, & Blondia, Chris. 2010. The missing ones: Key ingredients
towards effective ambient assisted living systems. Journal of ambient intelligence and smart
environments, 2(2).
Tharumarajah, A., Wells, A. J., & Nemes, L. 1996. Comparison of the bionic, fractal and holonic
manufacturing system concepts. International journal computer integrated manufacturing.
Tharumarajah, A., Wells, A.J., & Nemes, L. 1998 (Oct). Comparison of emerging manufacturing concepts.
Pages 325–331 of: Ieee international conference on systems, man, and cybernetics, vol. 1.
Valmero, A. 2011. E-health system to help curb maternal deaths in provinces Accessed on May 2,
2013 from http://ph.news.yahoo.com/e-health-system-help-curb-maternal-
deaths-provinces-054810305.html.
Vuidel, Gilles. 2013. Fractalyse — fractal analysis software.
Waring, J., McDonald, R., & Harrison, S. 2006. Safety and complexity: inter-departmental relationships as
a threat to patient safety in the operating department. Journal of health organization and management,
20(2–3) : 227–242.
Warnecke, H.J., & Hüser, M. 1993. The fractal company: a revolution in corporate culture. Springer.
Wolfram, Stephen. 2002. A new kind of science. Wolfram Media.
Ye, Juan, Dobson, Simon, & McKeever, Susan. 2012. Situation identification techniques in pervasive
computing: A review. Pervasive and mobile computing, 8(1) : 36–66.
Zhuge, Hai. 2010 (Nov.). Cyber physical society. Pages 1–8 of: Semantics knowledge and grid (skg), 2010
sixth international conference on.

Models and Concepts for Socio-technical Complex Systems: Towards Fractal Social Organizations

More Related Content

What's hot (20)

Similar to Models and Concepts for Socio-technical Complex Systems: Towards Fractal Social Organizations (20)

More from Vincenzo De Florio (20)

Recently uploaded (20)

Models and Concepts for Socio-technical Complex Systems: Towards Fractal Social Organizations