Available online at www.sciencedirect.

com

ScienceDirect
Procedia Computer Science 56 (2015) 566 572

International Workshop on Mobile Spatial Information Systems (MSIS 2015)

Integrating Multiple Geometric Representations within Spatial Data

Warehouse Structures for Enhanced Collaborative Decision-Making
Processes
Boubaker Boulekrouchea, Hedi Haddadb, Nafa Jabeurc*, Zaia Alimazighia
a
University of Sciences and Technologies, Houari Boumediene, BP 32 EL Alia, Bab Ezzouar Algiers, Algeria
b
Dhofar University, Computer Science Department, P.O. Box 2509, Post Code 211, Salalah, Sultanate of Oman
c
German University of Technology in Oman (GUtech), PO Box 1816, Athaibah PC 130, Muscat, Sultanate of Oman

Abstract

Spatial Decision Support Systems (SDSSs) have been successfully used to help decision-makers in several application domains.
When it comes to enabling the collaboration of all these decision-makers with one common SDSS, a Spatial Data Warehouse
(SDW) is commonly considered as an efficient option, particularly when large volumes of data are available. This SDW has to
accommodate all the decision-makers, anywhere and anytime, with the right geo-referenced data at the right time while respecting
the characteristics of their displaying devices. Since this aim is still unachievable due to the complexity of spatial processing, we
propose to shorten the processing time by leveraging the use of multiple representations of spatial objects in SDWs. For the sake
of illustration and unlike any existing solution, we focus in this paper on the integration of multiple geometric representations in
SDW and we propose new definitions to reflect the impact of this integration. We also highlight the importance of this integration
via the implementation of a collaborative scenario related to the management of industrial accidents in Algeria.

2015
2015 The
TheAuthors.
Authors.Published
Publishedbyby
Elsevier B.V.B.V.
Elsevier This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Conference Program Chairs.
Peer-review under responsibility of the Conference Program Chairs
Keywords:Spatial Data Warhouse; Multiple Geometric Representations; Collaborative Spatial Decision Support Systems

1. Introduction

Spatial events such as heavy rains and road accidents commonly have variable impact on the well-being of people
and their daily routine activities. The effects and the likelihood of these events are usually measured in every country

* Corresponding author. Tel.: +968-2206-1189

E-mail address: nafaa.jabeur@gutech.edu.om

1877-0509 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Peer-review under responsibility of the Conference Program Chairs
doi:10.1016/j.procs.2015.07.255
 Boubaker Boulekrouche et al. / Procedia Computer Science 56 (2015) 566 572 567

by several decision-makers in order to take the necessary precautions and make appropriate plans in case of hazards.
To this end, geo-referenced data on events of interest are stored in different types of databases depending on each
decision-makers requirements and perceptions. The volume of these databases is continuously increasing thanks to
the growing use of the rapidly expanding pervasive computing systems, including Wireless Sensor Networks (WSNs),
Radio-Frequency Identification (RFID) readers, radars, and satellites. To handle such large volumes of geo-referenced
data, Spatial Data Warehouses (SDWs) are commonly used. A SDW could be defined as a subject-oriented, integrated,
time-variant, and non-volatile database including a collection of both spatial and non-spatial data in support of
managements decision- making processes1. The navigation and analysis of the data could then be performed with
Spatial Online Analytical Processing (SOLAP) tools. For visualization purposes, Geographic Information Systems
(GISs) are widely used.
When it comes to operating on large volumes of data, human being reasoning capabilities are known to be limited.
Spatial Decision Support Systems (SDSSs) are then widely used to help decision-makers, especially in crisis situations.
Within a collaborative environment, a SDSS is expected to include a common SDW integrating all the data sources
owned by all the decision-makers. This SDW should include Multiple Representations (MR) of the phenomena of the
real world by capturing the decision-makers perceptions of the ongoing events, which constitutes an alternative to
ontology-based approaches. More specifically, MR allow the phenomena of interest to be described, for example, by
their geometries (the set of locations in the space), their semantics or descriptive data (name, entity type, etc.), and
their topological relations with other entities (inclusion, adjacency, etc.)2. In the literature, MRs could be associated
with three categories of multipliciy3: a geometric multiplicity (different shapes and positions), a graphic multiplicity
(different symbols), and a semantic multiplicity (different interpretations).In this paper we are interested in the
geometric multiplicity, which offers both vertical and horizontal navigation. Vertical navigation allows for changing
an object's geometrical representation from one resolution level to another with more or less details. This is called
multi-resolution MR. We talk about horizontal navigation when objects' geometrical representations vary according
to users points of view, at a fixed resolution. This is called uni-resolution MR. It has been proven that MR, and more
specifically Multiple Geometric Representations (MGR), have the ability to shorten the time-consuming process of
on-the-fly map generation, which is needed to provide decision-makers with personalized geo-referenced content4.
Several works5,6,7,8,9,10 have addressed the integration of MR in the multidimensional structures of SDWs. However,
in addition to failing in appropriately managing all the aspects of MR (geometric, graphic and semantic), no existing
work has explicitly supported the uni-resolution MR, which, we believe, is very important to accommodate several
decision-makers with appropriate contents as per the characteristics of their displaying devices (e.g., PC, PDA,
smartphone, etc.). We, therefore, argue that in order to generate the right geo-referenced data for the right decision-
maker, current SDSS solutions should extend their SDWs with appropriate mechanisms to conveniently support the
MR-related aspects. To this end, in this paper we propose a model and a notation that extent SDW with uni-resolution
MR capabilities, while focusing, for the sake of illustration, on multiple geometric representations.
In the reminder of the paper, Section 2 presents our model for the integration of multiple representations in SDWs.
To this end, it proposes some definitions and presents some new notations to reflect the support of multiple geometric
representations in SDWs. It also introduces our motivating scenario which is related to the collaborative analysis of
industrial accidents in Algerian cities. Section 3 presents our ongoing implementation of a prototype called GOLAP.
Section 4 discusses our work with respect to the state of the art. Finally, Section 5 concludes the paper and outlines
our future works.

2. Toward supporting multiple geometric representations in spatial data warehouses

2.1. Preliminary

SDWs store and provide access to large volumes of historical data. They are commonly structured around
multidimensional models (hypercubes). Within these models, several fundamental concepts are defined, mainly,
spatial facts, spatial measures, spatial dimensions, and spatial hierarchies. A spatial fact is considered as the subject
to be analyzed in a SDW. It may, for example, describe an event of interest (e.g., rain) that happened in the geographic
space10. A spatial fact contains a set of attributes with numerical values called measures (e.g., precipitation level). The
concept of spatial measure can be viewed in multiple ways, including: (i) a collection of pointers to spatial objects1,11,
568 Boubaker Boulekrouche et al. / Procedia Computer Science 56 (2015) 566 572

(ii) the result of spatial topological operators and/or metric operators12, or (iii) a member of a spatial dimension13.
Spatial dimensions are used to explore the measures from different analysis perspectives. A dimension is an abstract
concept for grouping data that share a common semantic meaning within the domain being modelled. They usually
contain hierarchies that allow to analyze detailed or aggregated measures using the drill-down and roll-up operations
of SOLAP systems. A spatial hierarchy represents an analytic perspective within the spatial dimension. It must include
at least a spatial level. A spatial hierarchy can be completely spatial if all its levels are spatial or partially spatial if it
includes at least one non-spatial level11.
The capabilities of SOLAP tools (as well as similar tools) to generate better customized geo-referenced contents
can be achieved by extending the SDWs models with Multiple Geometric Representations (MGR) of spatial data
according to several view-points and resolutions. This will particularly endow these tools with additional options to
provide decision-makers with an extended diversity of views as well as with more flexibility in representing and
navigating spatial data3.In what follows, we investigate the impact of integrating MGR in SDWs models. We
particularly propose new graphical notations to express this integration and revisit the definitions of the concepts of
spatial level, hierarchy, dimension, and measure.

2.2. A Model and a Notation for Representing Multiple Geometric Representations

Without entering into the debates about defining and modeling geographic phenomena, from a SDW perspective
we represent real-world phenomena as facts with different attributes and dimensions. In this paper we precisely focus
on representing spatial attributes of a real-world phenomenon (a fact) by the use of spatial objects. In this context, any
spatial object representing a real-world phenomenon may have multiple geometric representations associated to
different perceptions of the observers. A perception is often characterized by multiple criteria, including the users
point of view and profile, the geometric resolution, the scale, and the time14. In our case study, we particularly focus
on the user point of view pv and the geometric resolution r and thereby represent a perception p with p= (pv, r). In
order to describe the different geometric representations associated with a spatial object, we propose an approach based
on the notion of stamping14. To this end, we define below a new structure called Geometric Representations Structure
(GRS) (Figure 1.a) that allows the association of each spatial multidimensional feature with multiple geometric
representations. In this structure: (i) the p-value field is the stamping that corresponds to the perception value (i.e., all
the instances of GRS have the same perception value), (ii) the name field is an alphanumeric that refers to the name
that identifies the geometry of the spatial object associated to the perception value p-value, (iii) the geometry refers to
one of the geometric representations defined by the OGC (point, line, or polygon) associated to the perception value
p-value, and (iv) the spatial data type icon refers to the pictogram corresponding to the geometric representation. For
example, a spatial object firezone (referring to a given fire disaster) can be observed and stored according to four
perceptions, let us say P1(pv1, r1), P2(pv2, r1), P3(pv1, r2), and P4(pv2, r2), where: (1) pv1 and pv2 refer to an expert
in industrial accidents and an agent of civil protection respectively, and (2) r1 refers to the resolution used for scales
between 1:5.000 and 1:10.000 and r2 refers to the resolution used for scales between 1:10.000 and 1:50.000.
We use the GRS to build the main concepts of our model: Spatial Level supporting Multiple Geometric
Representations (SL-MGR), Spatial Hierarchy supporting Multiple Geometric Representations (SH-MGR), Spatial
Dimension supporting Multiple Geometric Representations (SD-MGR) and Spatial Measure supporting Multiple
Geometric Representations (SM-MGR) that we respectively define in what follows.

Definition 1: Spatial Level supporting Multiple Geometric Representations

A spatial attribute describes the spatial properties of a real-world phenomenon that takes place in a geographical
area15. The domain of this attribute is the set of values this attribute can take. A spatial attribute is said to be geometric
if its domain is a set of values of geometrical type, such as point, line, and polygon16. In our model, we consider that
spatial attributes can be associated to spatial dimensions, every spatial dimension includes at least one spatial
hierarchy18, and every spatial hierarchy is organized into different [spatial] detail levels. A level of a hierarchy of a
dimension is called geometric spatial level if it includes a geometric spatial attribute. This level is said to be Spatial
Level supporting Multiple Geometric Representations (SL-MGR) if its geometric attribute can be assigned to as many
geometries (geometric values) as perceptions.
 Boubaker Boulekrouche et al. / Procedia Computer Science 56 (2015) 566 572 569

Corporatr-Headquarter P-Value Perception value

Id-Corporate (b) Spatial data type icon

Geometry Name Geometric representation object identifier

#p
Geometry Geometric attribute

1 1 1
(a)
P1(pv1,r1) P2(pv2,r1) P3

Industrial_Accident AF(pv1) = NN
Name Name Name
Geometry
r Geometry
r Geometry
r Id-Dimension 1 AF(pv2) = DGU
(c)
Id-Dimension n
AF(pv3) = TGU
Number of deaths
P4(pv1,r2) P5 P6 Number of injuries
Geometry #p /AFs
Name Name Name
Geometry
r Geometry
r Geometry
r
1 1 1
P1(pv1,r1) P2(pv2,r1) P3
P7(pv1,r3) P8 P9

Name Name Name

Name Name Name Geometry Geometry Geometry
Geometry
r Geometry
r Geometry
r

Legend: Composition Association between SL-MGR and Geometric Representations Structure

Spatial data type icons of GRS

Fig. 1. (a) Geometric Representations Structure, (b) Spatial level supporting MGR, (c) Spatial measure supporting MGR

In order to represent graphically a SL-MGR, we propose the notation depicted in Figure 1.b. In our notation, the
GRSs which are stamped at different resolutions by the same point of view are grouped to form a cluster (that we call
cluster of geometric granularities) and are ordered from the most detailed resolution to the coarsest resolution within
that cluster (r1 is the coarsest resolution, r3 is the most detailed). When switching from one resolution to a coarser
resolution, the new geometry associated to a given spatial object could either be stored beforehand or obtained by
applying some map generalization operators17. The application of these operators must not result in the elimination of
geometric objects, particularly when these objects are important for the decision-making process and/or explicitly
requested to be displayed on the final map. Moreover, on the one hand, each cluster of geometric granularities
generates a set of multi-resolution geometric representations associated to a given spatial object. On the other hand,
the GRS stamped by the same resolution and associated to different points of views generates a set of single-resolution
geometric representations related to a given spatial object.
Definition 2: Spatial hierarchy supporting Multiple Geometric Representations
A spatial hierarchy is said to be geometric if all its levels are geometric spatial levels. Similarly, this hierarchy is
said to be Spatial Hierarchy supporting Multiple Geometric Representations (SH-MGR) if all its geometric spatial
levels are also supporting multiple geometric representations (i.e., they are SL-MGR).
Definition 3: Spatial Dimension supporting Multiple Geometric Representations
As already mentioned, a spatial dimension is a dimension that includes at least one spatial hierarchy18.This
dimension is said to be geometric if it includes at least one geometric spatial hierarchy. In the same way, the dimension
is said to be Spatial Dimension supporting Multiple Geometric Representations (SD-MGR) if it includes at least one
spatial hierarchy supporting multiple geometric representations (i.e., a SH-MGR).
Definition 4: Spatial Measure supporting Multiple Geometric Representations
Like in some existing works5,11, we model a spatial measure as a spatial attribute of a fact in the SDW. A geometric
spatial measure is defined as a spatial measure that is represented by a geometric spatial attribute15.This measure is
said to be Spatial Measure supporting Multiple Geometric Representations (SM-MGR) if it can be assigned as many
geometries as perceptions. Furthermore, as recommended in previous works5,19, spatial aggregation functions must be
defined when SM-MGR are used. Examples of such functions include Nearest-Neighbor (NN), Disjoint Geometric
Union (DGU), and Touches Geometric Union (TGU). We reflect these functions in Figure 1.c for the SM-MGR
geometry. To this end, as an example, we specify: (i) NN as an aggregation function for the set of perceptions {P1
(pv1,r1), P4 (pv1,r2), P7 (pv1,r3)}of the same point of view pv1; (ii) DGU as an aggregation function for the set of
perceptions {P2 (pv2,r1), P5 (pv2,r2), P8 (pv2,r3)} of the same point of view pv2; and (iii) TGU as an aggregation
function for the set of perceptions {P3 (pv3,r1), P6 (pv3,r2), P9 (pv3,r3)} of the same point of view pv3. In Figure 1.c,
AF(pvi) refers to a spatial aggregation function of SM-MGR associated to point of view pvi. For simplification
purposes we don't illustrate the whole clusters in Figure 1.c.
570 Boubaker Boulekrouche et al. / Procedia Computer Science 56 (2015) 566 572

2.3. Motivating Scenario

Wilaya Industrial_Zone Corporatr-Headquarter
Id-Zone Id-Corporate
Id-Corporate
Other attributes Other attributes
Geometry #p Geometry #p Geometry #p

1 1 1 1 1 1 1 1 1
(pv1,*) (pv2,*) (pv3,*) (pv1,*) (pv2,*) (pv3,*) (pv1,*) (pv2,*) (pv3,*)

Name Name Name Name Name Name Name Name Name

Geometry Geometry Geometry Geometry Geometry Geometry Geometry Geometry Geometry

Industrial_Accident
All-Type Type
Id-Dimension 1
Id-T ype . Cause
Others
Id-Dimension n Month Id-C ause
Others
Number of deaths month
1 1 1 Number of injuries
(pv1,*) (pv2,*) (pv3,*) Geometry #p
All_Causes
Name Name Year
Name
Geometry Geometry Geometry Year

Fig. 2. Representing the motivating scenario of industrial accidents in Algeria

Several decision-makers (stakeholders) in Algeria are currently collaborating in the management of industrial
accidents in the different cities of the country. Each of these stakeholders, including industrial risk experts, civil
protection managers, and public safety officers, has its own specific risk evaluation, perception, and map-based
geometric representation. These differences commonly increase the complexity of risk analysis and make
collaboration almost impossible. There is therefore an urgent need for an efficient supporting tool where MGR
can be stored and generated on-the-fly or on-demand for each stakeholder. We argue that instead of using several
SDWs, a single SDW-based application using the same schema and the same statistical data (thematic) - but
supporting MGR related to risk localization - would allow the different stakeholders to explore spatial data
simultaneously, according to multiple perceptions. This will consequently boost their collaboration and leverage
the tedious and complex data analysis and decision-making processes. Our aim is thus to provide stakeholders with a
system for spatial multidimensional analysis of data on industrial risks. This system should allow for the analysis of
the numbers of deaths, number of injuries depending on the type of risk, the cause of accidents, and the
geographical location and time of accidents. We depict in Figure 2 a possible design of the SDW supporting MGR
for the scenario of monitoring of industrial accidents.

3. Implementation

For the sake of illustration, we developed a SDSS called GOLAP to provide on-the-fly personalized geo-decisional
information to different decision-makers according to their preferences and to the visualization characteristics of their
terminals (desktop, PDA, mobile phone, etc.). GOLAP system was implemented using ROLAP server Mondrian and
the PostGIS spatial DBMS for the processing and analysis of non-spatial and spatial data respectively. We used a star
schema to model our SDW. Figure 3.a shows our relational data model used in the implementation. The fact table used
is Industrial_accidents. It includes four foreign keys to the dimensions reflecting the cause of the accident, its type,
the zone where it occurred, and the time when it occurred. It also includes two measures nbdeaths and nbinjuries as
well as the SM-MGR localization.
In our relational model, we propose to associate each spatial level of a given spatial dimension to a new table
(representation table) that represents the structure holding multiple geometric representations. Each of these
representation tables (e.g., Local_Rep in Figure 3.a) contains fields to represent its primary key (e.g., Local_repPK in
Figure 3.a), a foreign key from a geometric table (e.g., LocGeoFK in Figure 3.a) and a perception attribute. Each
geometric table contains fields to represent its primary key as well as to store its geometrical representation
(Geometry). It is important to remark that without the geometric table, a spatial dimension with spatial data would
have to store all multiple geometries together, which will not be cost-effective in terms of storage space.
 Boubaker Boulekrouche et al. / Procedia Computer Science 56 (2015) 566 572 571

Corp-Geo Local_Rep Time <?xml version= "1.0"? >

CorGeoFK
Local_repPK
LocGeoFK
id-Month
Month
-
-
<Schema name= "industrial_accidents"">
<Cube name= "indust_accidents"> (b)
Geometry
Perception Year <Table name= " indust_accidents" >
+ <Dimension name= "Time" foreignKey=Idmonth >
+ <Dimension name= "Cause" foreignKey=Idcause >
+ <Dimension name= "Type" foreignKey=Idtype >
Industrial_Accidents + <Dimension name= "Zone" foreignKey=IdZone >
Type - <Hierarchy hasAll="true" primaryKey="zone_id" primaryKeyTable="Zone">
Id-Zone <Join leftKey=" Corp_repFK," rightAlias=" Corp_Rep" rightKey=" Corp_repPK ">
Id-time
(a) Id-Cause
id-Type
Type
<Table name="Zone"/>
<Join leftKey=" CorGeoFK" rightKey=" CorGeoPK ">
Id-Type <Table name=" Corp_Rep "/>
<Table name=" Corp_geo "/>
nbdeaths ...
nbinjuries
Cause
</Join>
Localization <Level name= " Corporate" column= " Corporate" uniqueMembers=" true">
id-Cause
Cause <Property name=" IdCorp_rep" column= " Corp_repFK" />
< /Level>
Zone <Level name= "Corp_representation" uniqueMembers=" true">
id-Zone Indu_Rep <Property name=" Idcorp_rep " column= " Corp_repPK" />
Wilaya_Rep
Corporate Indu_repPK <Property name=" Idgeo_corp" column= " CorGeoFK" />
Corp_repFK IndGeoFK Wil_repPK <Property name=" PERCEPTION" column= " Perception " />
Indu_zone Perception WilGeoFK < /Level>
Indu_repFK Perception <Level name= " Corp_geometry" uniqueMembers=" true">
Wilaya <Property name=" idcorp_geo " column= " CorGeoPK" />
Wil_repFK <Property name=" GEOMETRY" column= " Geometry" />
< /Level>
Indu_Geo Wilaya-Geo
Loc_geo < /Dimension>
IndGeoFK WilGeoFK </Cube>
LocGeoPK Geometry Geometry
Geometry </Schema >

Fig. 3. (a) Relational model of our MGR of our motivating scenario, (b) Excerpt of the Mondrian XML schema
In order to implement our model, we depict in Figure 3.b an excerpt of the Mondrian XML schema that we used
for our motivating scenario. Currently, the client side of our prototype is using OLAP client Jpivot for the visualization
of non-spatial data and the JavaScript OpenLayers tool for the visualization of spatial data. A special Processing and
Personalization Module of our prototype is in charge of customizing the content of the final map in order to meet the
constraints of different displaying devices. This module, being developed in Java, is also responsible for selecting
and/or generating the appropriate MGR for the current service required by a given decision-maker.

4. Discussion and related works

The main contribution of our work is the introduction of a new model and a notation for the integration of uni-
resolution multiple geometric representations within SDWs. Despite its important contribution to shorten the
collaborative decision-making processes and the generation of personalized map contents, only few works have
integrated MR within the SDWs design and implementation. To the best of our knowledge, these research works do
not appropriately support the uni-resolution MR concept as we presented in this paper. In this regard, Bernier and
Bedard5 and Bedard et al.6 have used the concept of Vuel (supporting geometric, semantic and graphic multiplicities)
to represent spatial objects in a SDW for the generation of maps on-demand. However, both proposed solutions did
not explain how MR are actually supported at different levels of the multidimensional structure of the
SDW.Bernier7hasdefined the MR of spatial objects as pictograms to which he applied spatial processing techniques
to generate the required maps. Damiani and Spaccapietra8have described a conceptual model called MuSD
(Multigranular SDW) where semantic and graphic multiplicities are supported in a multi-resolution context. MuSD
adopts a spatial data model based on OGC standards and includes additional SOLAP operators to navigate in several
geometric granularities. However, in addition to not supporting multiple geometric representations, the processing of
geo-referenced data was not addressed at the uni- resolution level. Bimonte et al.9 have presented and extended a
SOLAP model along with an algebra that supports map generation. The proposed approach is capable of generating
new relations (called multi-association) between spatial objects but does not support the geometric multiplicity of
these objects in single-resolution. A more detailed discussion and review of the state of the art is out the scope of this
paper and will be covered in other publications.

5. Conclusion and future works

In this paper, we investigated the importance of enabling and facilitating the collaboration of several decision-
makers on some ongoing events of interest based on Spatial Decision Support Systems (SDSSs). We emphasized the
572 Boubaker Boulekrouche et al. / Procedia Computer Science 56 (2015) 566 572

need to leverage the integration of multiple representations of spatial objects and events in Spatial Data Warehouses
(SDWs) when large volumes of data are used. To this end, and unlike existing solutions which did not adequately
support the multiple representations issues, especially at the uni-resolution level, we proposed to focus on the
integration of Multiple Geometric Representations (MGR) in the SDWs. We presented new notations to support the
modelling of such SDWs and extended the definitions of some fundamental properties related to the multidimensional
structure of these warehouses. We applied our concept to the scenario of monitoring industrial accidents in Algeria
and reported some implementation details.
Our ongoing works are focusing on the definition and implementation of a new algebra supporting data navigation
across MGR. We are then planning to explore the challenges to Extract, Transform, and Load spatial data in a MGR
context.

References

1. Stefanovic, N. Han, J. and Koperski, K., 2000, Object-Based Selective Materialization for Efficient implementation of Spatial
Data Cubes. IEEE Transactions on Knowledge and Data Engineering, Vol. 12, n 6, pp. 938-958.
2. Devogele, Thomas, 1997, Processus d'intgration et d'appariement de Bases de Donnes Gographiques: Application une base de
donnes routires mulri-chelles, Thse de Doctorat de Mthodes Informatiques, Universit de Versailles, 207 pp.
3. Martel, C., 1999, Dveloppement dun cadre thorique pour la gestion des reprsentations multiples dans les bases de donnes
spatiales, Mmoire de matrise, Universit Laval,
4. Ruas, A. and Bianchin, A., 2002, Echelle et niveau de dtail, dans : Ruas A. (sous la direction de), Gnralisation et reprsentation multiple,
Paris, Herms Science Publications, , pp. 25-44.
5. Bernier, E., and Bdard, Y., 2007, A data warehouse strategy for on-demand multiscale mapping. In Mackaness, W. A., Ruas,
A., Sarjakoski, L. T. (Eds.), Generalisation of geographic information: Cartographic modelling and applications Amsterdam, The
Netherlands, pp. 177198 (Elsevier Science).
6. Bdard, Y., Proulx, M.-J., Larrive, S. and Bernier, E., 2002, Modeling multiple representations into spatial datawarehouses: A
UML-based approach. In Proceedings of the Joint Workshop on Multi-Scale Representations of SpatialData, ISPRS WGIV/3, ICA
Commission on map Generalization, Ottawa, ON, Canada.
7. Bernier, E., 2002, Utilisation de la reprsentation multiple comme support la gnration de vues de bases de donnes gospatiales
dans un contexte SOLAP Mmoire de M. Sc., Facult de foresterie et de gomatique, Universit Laval, 2002
8. Damiani, M. and Spaccapietra, S., 2006, Spatial data warehouse modeling. In Darmont, J., and Boussaid, O. (Eds.), Processing and managing
complex data fordecision support, pp. 126 (Hershey, PA: Idea Group).
9. Bimonte, S., Bertolotto, M., Gensel, J. Boussaid, O., 2012, Spatial OLAP and Map Generalization: Model and Algebra, International Journal
of Data Warehousing and Mining, January-March, 8(1), pp 24-51.
10. Badard, T., Bdard, Y., Hubert, F., Bernier, . Dub, . (2008) Web services-oriented architectures for mobile SOLAP applications,
Int. J. Web Engineering and Technology, Vol. 4, No. 4, pp.434464.
11. Malinowski E. and Zimnyi E., 2004, Representing spatiality in a conceptual multidimensional model. In : PFOSER Dieter,
CRUZ Isabel F. et RONTHALER Marc. 12th ACM International Workshop on Geographic Information Systems, 12-13 Novembre,
Washington, USA. New York, USA, pp 12-22 (ACM Press, 2004).
12. Rivest, S., Y. Bdard, M.J. Proulx, M. Nadeau, F. Hubert and J. Pastor. (2005) SOLAP: Merging Business Intelligence with Geospatial
Technology for Interactive Spatio-temporal Exploration and Analysis of Data. Journal of International Society for Photogrammetry and
Remote Sensing (ISPRS), 60(1),17-33.
13. Marchand, P., A. Brisebois, Y. Bdard et G. Edwards (2003). Implementation and evaluation of a hypercube-based method for spatio-temporal
exploration and analysis. Elsevier: Journal of the International Society of Photogrammetry and Remote Sensing, Vol. 59, n 1, 6-20.
14. Parent, C., Spaccapietra, S. and Zimanyi, E., 1999, Spatio-temporal conceptual models: Data structure + space + time, in Actes de la
confrence GIS99, Kansas City, Missouri (USA).
15. Bimonte, S., 2007, Intgration de l'information gographique dans les entrepts de donnes et l'analyse en ligne: de la modlisation
la visualization, PhD Thesis, INSA, Lyon, France, 207 pages.
16. Da Silva, J. de Oliveira, A.G., Fidalgo, R. N., Salgado, A.C., Times, V.C. (2009). Modelling and querying geographical data warehouses,
in Information Systems journal.
17. Gascuea, C., and Guadalupe, R., 2009, A multidimensional methodology with support for spatio-temporal multigranularity in the conceptual
and logical phases. In Taniar, D. (Ed.), Progressive methods in data warehousing and business intelligence: Concepts and competitive
analytics, pp. 194230 (Hershey, PA: IGI Global).
18. Malinowski, E., and Zimnyi, E., 2008, Advanced data warehouse design from conventional to spatial and temporal applications.
Berlin, Germany: Springer-Verlag
19. Bdard, Y. and Han, J., 2009, Fundamentals of Spatial Data Warehousing for Geographic Knowledge Discovery. Geographic
Data Mining and Knowledge Discovery (2nd edition), H.J. Miller and J. Han (Taylor & Francis).