HYPER-MSPACE; MULTIDIMENSIONAL EVOLUTIONARY-AGENTS MODELING AND ANALYSIS

International Journal of Computer Science & Information Technology (IJCSIT) Vol 5, No 3, June 2013
DOI : 10.5121/ijcsit.2013.5303 35
HYPER-MSPACE; MULTIDIMENSIONAL
EVOLUTIONARY-AGENTS MODELING AND ANALYSIS
Mohamed Dbouk1
, Ihab Sbeity1
, Hamid Mcheick2
, Haytham Douaihy3
1
Lebanese University, Faculty of Sciences (I), Beirut-Lebanon
mdbouk@ul.edu.lb, isbeity@ul.edu.lb
2
Université du Québec à Chicoutimi, Québec, Canada
Hamid_Mcheick@uqac.ca
3
Lebanese University-DSST, Beirut-Lebanon, Master2R-Student
haythamdouaihy@gmail.com
ABSTRACT
This paper stands for an analytical approach modelling and exploring an evolutionary-space oriented
business-domains. The space that we consider is intended to be complex emerging a set of heterogeneous
evolutionary business-agents (e.g. Factories, Restaurants, etc.); it accordingly stimulates an analytical
multi-layered structure headed for a hyper-space navigational framework. Agents have multiple-states
based on analytical characteristics (analytical dimensions), they are multidimensional and inter-
dependable; they evolve, interact and transit. This approach aims to provide key features that illustrate the
hyper-space framework basis and foundations. Finally, an instructive case study is given at the end of this
document.
KEYWORDS
Evolutionary agents, Hyper-space modelling, Multi-State modelling, GIS-like frameworks, Business
intelligence.
1. INTRODUCTION
Day by day, reactive and dynamic data are strongly solicited. This data may incorporate and
stimulate gigantic amount of knowledge. The advance in technology, multidimensional and
business intelligence analysis are extremely present in similar context.
However, things evolve in the nature, everything evolve and change states and conditions, e.g. an
employee may become manager, president, etc., modern cities, construction sites, volcanic sites,
etc., may have continuous stories and histories (images, video, maps, etc.); they evolve
continually against time, urbanisations and civilizations.
Concepts of evolutionary space may, indeed, reside in lots of real-world events and business
activities; environmental risk related issues such as flooding, forest (trees, etc.) evolutions,
agricultural practices, are typical cases.
In our early MSpace platform [1], all considered agents are belong the same type, i.e. employers,
buildings, etc. Moreover, the space itself is reduced to one and only one business
context/environment (layer).
The ultimate goal of this work is to increase and enrich our MSpace platform with new hyper-
space characteristics and functionalities; the space is multi-layered now and become so complex

36
by emerging heterogeneous cooperative and evolutionary agents. However, dealing with such
unconventional space architecture/structure is hard to undertake and reveal great challenges. The
multi-layering concept leads to provide an intelligent and analytical vertical navigation.
The duty is to ultimately devise, on one hand, a strong and firm theory correlating the intended
hierarchical structure; towards an analytical-oriented specialization/generalization strategy that
incorporates the vertical navigational and analytical transitions. The duty, on the other hand, is to
correlate interactions and behaviours of the considered heterogeneous agents each other, and also
to adjust the inter/intra-events which trigger transitional evolutions.
The rest of the paper is organized as follow: in Section 2 we review background material, the
early MSpace as well as problem statement are outlined and an overview of related works is also
given in this section. In section 3, MSpace extension foundations and basis are deeply
considered. Also, a deep modelling theory and navigation as well as visual solution are presented
in this section. An experimental case study is also given in section 4. Finally, the outlines of
future directions and conclusion are drawn in section 5.
2. BACKGROUND AND PROBLEM STATEMENT
2.1. MSpace, an overview
The early MSpace [1] (started with [2]) consists of an integrated tool tending to model and
explore (tracking and monitoring) an evolutive and mutable space, it deals and emerges
multidimensional and unconventional data, it mainly considers a collection of mutable entities
(became agents), which are referenced by a structural features (dimensions), and tracked
according to behavioural features (transitions, business rules).
Entities/Agents in MSpace are expected to move, transit, navigate, interact, etc. agents’
behaviours reflect and/or materialize external events. Rather than temporal and location-based
characteristics, they evolve in well organized behaviours. MSpace deals with contextual patterns
(business profiles, figure 1), in which dimensions are pre-combined.
Figure 1. MSpace [1] platform basis and foundations
The ultimate goal, of “MSpace”, is to enable entity (individual space’s element) evolution
tracking and monitoring. The duty is to predict, direct and ultimately correlate entities’
behaviours and activities.
Considered
Agents
Business named dimensions
MSpace
Business Profiles
(Total-States)
A named profile
Time
EducationYears of duty
Manager
General-
Secretary
Analyst
Controller
…
Case-study: Human Resource Monitor
Profile examples

37
2.2. Problem position and related works
In the practice, business-spaces are usually complex and considered entities/agents could
reference different categories and classes, they may interact and influence each other, etc.
Consider the typical case of a modern city evolution analysis. To analyze and track the evolution
of the city, we have, indeed, to record and analyze the urbanization activities from multiple
perspectives; i.e. the existence of a Factory may infect healthy conditions, the existence of
buildings requires the presence of streets, the existence of 20 buildings or more requires the
presence of gardens, etc.
Here the space may contain heterogeneous and inter-dependable agents; it is complex and hard to
undertake. The concept of subdividing the space into bordering areas/subspaces and or applying
the multi-layering approach may be helpful, but dealing with such cases requires deep modelling
efforts. The idea, indeed, behind this case is to exhibit the big challenges that we are facing
where the space is qualified as complex.
However, in the early above approach [1], complex cases are not yet solved. Rather than the
elegant and hopeful outcomes, the experiments show that we need to deploy more efforts in order
to enlarge the analytical spectre of this approach.
Obviously, dealing with evolutionary-objects or agents is sometimes synonymous to deals with
moving objects or agents on one hand and to consider the spatiotemporal dimension/concept on
the other hand.
– The FuMMO model proposed by [5] works indeed with geo-features, the moving objects
data model “Balloon” [6] deals with spatiotemporal changes; it provides an integrated
support for both historical and future movements of moving objects. In contrast to basic
“FuMMO” model, “Balloon”‎states for emerging both historical and future moving-objects’
movements.
– “Hermes” [4] deals with dynamic objects that change locations, shape and size, either
discretely or continuously over time.
– In the land use model presented in [3], the authors use discrete irregular objects as an
autonomous spatial entity. They use the basic elements of Cellular Automata: space,
neighbourhood, and transition rules, state, and time, and demonstrate that the proposal can
be considered as a new paradigm for urban simulation. Here, the world is considered as a
series of entities located in space. An autonomous active object or agent is a digital
representation of all part of an entity. In this model, the transition rules specify the
behaviour of cells between time-step evolutions, deciding the future conditions of cells
based on a set of fixed rules. However, the state of the object changes is based on the
adaptation of an entity to its environment. Also the time variant is considered as an element
in this approach.
Briefly, we plan, in the next sections, to extend the early MSpace approach by adding the so
called hyper-features, and functionalities.

38
3. HYPER-MSPACE
3.1. Complex space modelling
The new-extended initiative “Hyper-MSpace” emerges indeed features and concepts which are
considered by the above premature approaches ([3], [4], [5] and [6]), it inherits concepts from
multi-agents systems, but remaining different on the way to define agents.
Here, we are going to devise a deep theory leading to model that we call “Hyper-MSpace”; rules
that govern agents interactions, space subdivision rules, multi-layering space constitution,
vertical navigational rules (summarising, zooming, etc.), etc. As a result, we state here for a
complete complex-space building and exploring process; we consider answering questions like:
“how the space should be partitioned thematically”, “how entities could interact and influence
each other”, “how could navigate through different levels”, etc.
Basic assumptions: Hyper-MSpace (HMSpace) consists of an integrated space, it emerges
several agent categories, consequently it incorporates a set of MSpaces (figure 2)
Formally, HMSpace (C1, C2, …., Ci, …Cn) = <MS1, MS2, …, MSi, …, MSn>
– MSi = MS(Ci); MSpace (figure 1) is incorporating entities of type Ci.
Consequently, it should instantiate all mspace pre-elaborated features.
– Ci is an agent multidimensional type
Figure 2. HMSpace basic constitutions
Inter-dependability modelling:
Emerged agents in HMSpace are intended to influence each other. An evolutionary agent (also
called receiver) may depend on other agent/s (called master). Accordingly, an oriented mutual
association is so established between master-agents and receiver-agents. Consequently, a huge
inter-agents dependability network could be established, it represents a derivation /instantiation
∆: MSpace of ∆
: MSpace of 
: MSpace of 
Complex/Composite Space
– incorporates several
entity/agent categories,
– integrates set of MSpaces
Inter-MSpace
dependability

39
of an extrapolated network called “inter-agents dependability diagram (IADD)” (figure 3). An
IADD consists of an oriented graph; nodes represent the involved mspaces and the arrows
represent the inter-dependence relationships between agent-instances.
However, when a master-agent “” or “∆” evolve, it could affect (influence) all dependent
receiver-agents “” (figure 4). Based on the mspace firm foundations, the receiver-agent evolve
throughout dimensions, consequently:
– A master-agent affects receiver-agents on one of their pre-defined dimensions. Arrows in
dependability-diagram should be labeled with the given receiver-agent dimension.
– Master-agent may affect receiver-agents on many dimensions.
– Master and receiver agents may share (be belong) same mspace, towards reflexive
dependability.
Figure 3. Inter-mspaces dependencies diagram
Figure 4.a. HMSpace platform basis and foundations
Materializing inter-dependability:
The association that links master-agents and receiver-agents can be materialized using the
Observer/Observable design pattern; it is event-driven based.
∆


An instance
There are some agents from, that influence the agent
from.
ID diagram
, , and ∆ are three different
mspaces
∆

 /D3
/D2
/D1
/D1

Dependability
in depth

∆
Business
dimensions
Business
profiles
Dependence
relationship

40
Here, receiver-agents play the role of observers, and variations or evolutions on one master-agent
are triggered. Consequently, local-transitions from state to state of a receiver-agent (on the given
dimension like D3, figure 4) are now directly linked to the master agent’s total-transitions
(locally tracked in the given mspace; business-profiles).
Formally;
∆ is a master-agent / observable pre-declared in mspace ∆ (e.g. ∆: ∆)
 is a receiver-agent / observer pre-declared in mspace  (e.g. :)
Protocol:
1. ∆ Transits form local-state to a new state, consequently, a new total-state should be
generated in ∆.
2. The new ∆ state is observed by  and triggered.
3.  applies a dedicated transition-function “.Dim”, it analyses the ∆ new state
and decides to transit from its old state to another local state on the given
dimension, it should generate a new total-state in , and then new business-profile
may be generated as well.
Figure 4.b. HMSpace platform basis and foundations
Transition-function formulation and generalization:
.D3(∆); transition of agent, on dimension “D3” taking into account the ∆ evolutions.
More general:
– may be concerned with more than one ∆;.D3(∆1,∆2,…, ∆i,…, ∆n)
– may be linked to more than one agent-mspaces;here and aggregation “” must be
applied: .(Di(…) ).
However, those dependability functions are domain-dependant and should be written by the
domain analytical experts. As a result, the HMSpace platform has become open.
Dynamic aspect:
According to the pre-elaborated IADD (figure 3), consider that agent 4 depends on {<∆1, ∆3,
∆5>, <1, 4, 6>} , and the new agent ⌂2 depends, in its turn, on 4, the following scenarios
should be considered:
– On cascade-evolution; if one of the {<∆1, ∆3, ∆5>, <1, 4, 6>} evolves, the 4 should be
mspace
∆: ∆
mspace
: D3
(1)
(1)
(2)
(3)
(3)


41
affected and may change state. Consequently, ⌂2 will (by transitivity) evolve as well, and
so on. All related agents will evolve, toward an on cascade-evolution.
– Dependability-pertinence: agents evolve continually, after each evolution-round, the
IADD instances should be re-examined and evaluated. As a result, 4 for example may
lose (become free of) some links, and also new links may re-appear with respect to the
pre-elaborated IADD, i.e. 4  {<∆1, ∆5>, <1, 5, 6>}.
Visual and navigation issues
The early mspace platform [1] consists of a set of pre-elaborated visual and analytical operations.
Here in the HMSpace context and based on the given IADD, each mspace is represented as an
analytical-node, and the inter-dependability links represent the firm foundation of two additional
operations:
– Navigation throughout mspaces: allowing end-users to explore the whole considered
space (hmspace), and to discover mspaces one by one, this is to materialize relationships
“master-receiver” influences, e.g. ∆1 4
– Analytical zooming: this is to explore a given agent or mspace and going forward on a
specific dimension. Here, the end-user must be able to discover and learn about the
transition-modalities.
Figure 5. HMSpace end-user display basis and foundations
As a result, the above operations are complementary to each other. The IADD is, continually,
considered as the main display that summarizes hyper-space new essential operations.


∆
Analytical mspace
environment
⌂
Navigation, based on dependability-link, from an mspace to another
Analytical zooming, based on dependability-link, from an mspace to another
Dependability-
link

42
3.2. Hyper-Space building process
The building process consists of the following:
a. Identify and consider the involved agent-mspaces and all descriptive and analytical
features or dimensions,
b. analyze and find-out the inter-dependability relationships that relate the above mspaces,
c. Put mspaces and relationships together in a unique IADD,
d. re-adjust the above mspaces according to the founded relationships,
e. designate labels that governing the above dependability relationships,
f. Draw up (pseudo code) the transition-functions one by one.
g. Implementation phase: put the given IADD into operational platform.
h. Apply analytical and predictable scenarios.
Practically, the process is software-engineering-like, it requires key persons to be applied, and
they should emerge technical-skills as well as deep knowledge in the considered business-
domain.
4. EXPERIMENTAL CASE STUDY AND OBSERVATION
Consider the case of patients tracking and monitoring. Here, the main challenge is to produce and
implement the equivalent IADD.
However, we are going to focus and deal with patients (figure 6) and their analytical dimensions;
administered-medicines and properties, treatment modalities, body building issues, climate
issues, residency (home-residence), etc.
Figure 6. Patients tracking case study
Body-Conditions
Patients
FactoriesAgriculture
Zones
Medical Supports
Medical
Laboratories
Sport-Centers
Applied-Medicines
Medical-Treatment
…
Body-Conditions
…
Geo-Location
…
Climate-Conditions
Infection-factors
…
Patient-Profiles
Medical-
Treatment
Climate-
Conditions

43
If, for example, patients are concerned with the climate-variations and pollution degrees, then we
have to find out the concrete agents that influencing the climate-conditions like “Agriculture-
Zones” and “Factories” (figure 6). The same conceptual analysis leads to consider “Medical-
Supports” and “Medical-Laboratories”, “Sport-Centers”, etc., each of which relates (labeled
with) one analytical dimension.
To conclude, first conceptual experiments show that, extra-agents could be grouped/aggregated
into some abstract–agent-types, this is because those extra-agents share (influence) same
analytical-dimensions, e.g. “Climate-agent” may aggregate “Agriculture-Zones” and
“Factories”. Finally, the experiments also show that we are pressed to deploy additional efforts
that lead to implement the above theory and concretely validate the delivered process.
5. CONCLUSION
The early “MSpace” approach that we proposed in [1] provides a flexible modeling and
analytical framework, but it deals and considers only one agents type. However, extending this
approach represented, for us, the big challenge.
We stated, during the above work, for a complex-space called HMSpace (Hyper MSpace) that
emerges heterogeneous agent-types and deals with inter-agents interactions. As a result, we
divested a deep inter-agents dependability theory which was formulated with a new elegant
UML-like structure called IADD (Inter-Agents Dependability Diagram).
Rather than the hopeful outcomes, the experiments show that we need to deploy more efforts at
the validation phase. Finally, extending the early MSpace software-prototype outlines our future
research.
ACKNOWLEDGEMENTS
The authors gratefully acknowledge support from the Lebanese University Research
administration, Doctoral School of Sciences & Technologies (EDST) – Beirut – Lebanon
REFERENCES
[1] Dbouk, M., Mcheick, M. & Sbeity,I. (2012) "MSpace: An Emergent Approach, Modelling and
Exploring Mutable Business-Entities", Advances and Applications in Mobile computing, Publisher
InTech, Edited by Adem Karahoca, Chapter 6, pp 115-136.
[2] Dbouk, M. & Sbeity, I. (2009) “GeoSEMA: A Modelling Platform, Emerging GeoSpatial-based
Evolutionary and Mobile Agents. World Congress on Science, Engineering and Technology (WCSET
2009); Conf. Proceedings ISSN: 2070-3740, March 23-25, 2009, Hong Kong, China.
[3] Hammam, Y., Moore A., Whigham P. & Freeman C., (2004) “Irregular vector-agent based
simulation for land-use modelling”. In 16th Annual Colloquium of the Spatial Information Research
Centre (SIRC 2004: A Spatio-temporal Workshop), 29-30 November 2004, Dunedin, New Zealand,
pp. 103-116.
[4] N. Pelekis, Y. Theodoridis, S. Vosinakis, T. Panayiotopoulos, “Hermes - A Framework for Location-
Based Data Management”, EDBT 2006, Springer Lecture Notes in Computer Science, Vol. 3896,
pp.1130-1134.
[5] Praing, R. & Schneider M. (2007). A Universal Abstract Model for Future Movements of Moving
Objects. In AGILE International Conference on Geographical Information Systems, 2007.

44
[6] Praing, R. & Schneider M. (2007). Modeling historical and future movements of Spatio-temporal
objects in moving objects databases. In proceedings of the 16th ACM Conf. on Information and
Knowledge Management (CIKM), p183-192, Lisbon- Portugal, Nov. 6-10, 2007.
AUTHORS
Mohamed Dbouk, the main author, is a full time Professor at the Lebanese University
(Beirut-Lebanon), he coordinates a master-2 research degree “IDSS: Information and
Decision Support Systems”. He received his PhD from “Paris-Sud”, France, 1997. His
research interests include software engineering and information systems related issues,
geographic information systems, data-warehousing and data-mining.
Ihab Sbeity is a full time Associative Professor at the Lebanese University – Faculty of
Sciences I –Department of Computer Science. He is member of the IDSS teaching-team.
He received a PhD from the “Institut National Polytechnique de Grenoble”, France. His
PhD works are related to Performance Evaluation and System Design. His research
interests include software performance engineering, UML modeling, multi-criteria
decision making and information systems.
Hamid Mcheick is currently full time Associate Professor in Computer Science
Department at the University of Quebec at Chicoutimi (UQAC), Canada. He received a
Master’s degree and PhD in Software Engineering and Distributed System from Montreal
University, Canada. He is interested in software development and architecture for
enterprise applications as well as in separation of concerns (component, services, aspect,
etc.). His research is supported by many research grants he has received from the
Canadian Government, University of Montreal, “Centre de Recherche informatique de
Montreal (CRIM)”, University of UQAM, and University of UQAC.
Haytham Douaihy attended the Graduate School of Sciences & Technology in the
Lebanese University, for earning a Master degree Research in Information Systems. In
October 2007, He started working at “Path-Solutions”, a Leading Software Provider
company in the Banking Industry, In June 2011; he got a position as a Senior Web
Developer at a Software Provider company in the Banking Industry: 'Xperteam'.

HYPER-MSPACE; MULTIDIMENSIONAL EVOLUTIONARY-AGENTS MODELING AND ANALYSIS

More Related Content

Viewers also liked (9)

Similar to HYPER-MSPACE; MULTIDIMENSIONAL EVOLUTIONARY-AGENTS MODELING AND ANALYSIS (20)

Recently uploaded (20)

HYPER-MSPACE; MULTIDIMENSIONAL EVOLUTIONARY-AGENTS MODELING AND ANALYSIS