CARTOGRAPHIC EXPERT SYSTEM FOR SUSTAINABLE DEVELOPMENT OF

THE ADMINISTRATIVE TERRITORIAL UNITS (ATU)

Ioan STOIAN*, Vasile NACU *, Dan VELE**

*Centrul Național de Cartografie – BUCURESTI
**Facultatea de Geografie - UBB CLUJ NAPOCA

This article briefly presents only the steps used to make a cartographic expert system
for sustainable development in order to create the necessary maps, used in risks management
and hazards caused by floods, landslides and earthquakes, which incorporates: statistical
analysis of experimental data, optimization in decision-making processes modelling, planning
and network control of complex projects, simulation of complex decision-making processes,
etc.

1. INTRODUCTION

Starting from the well-known definition of sustainable development given by the

World Commission on Environment and Development (WCED) -1983-in the report "Our
Common Future", also known as the Brundtland Report: "sustainable development is
development that aims to meet the needs of the present , without compromising the ability of
future generations to meet their own needs.”
The sustainable development follows and tries to find a stable theoretical framework in order
to make decisions in any situation where a human-environment report is found, even if it`s
about economic or social environment.
A number of 17 Sustainable Development Goals (SDG), or the Global Goals assumed by our
country, along with the other 193 states members of UN in September 2015 cover a wide
range of themes that promote global action in three main areas of sustainable development:
economy, society and environment. The 17 SDG`s are equally addressed to both,
underdeveloped and developed countries and regions. Thus, by 2030, the world's states have
committed themselves to eliminate poverty and hunger, fight inequality and injustice, and
take active measures to protect the environment (see Fig. 1).Regarding this, efforts are mainly
focused on:eradication of poverty; changing patterns of production and consumption;
eradication of poverty;changing patterns of production and consumption; protecting
health and Protecting and managing the base of natural resources for economic and
social development

Fig.1 Sustainable Development Goals (SDG)

Local authorities must react to both the requirements of the population for protection against
natural calamities as well as to the reduction of material damage caused by them, as well as
alarm signals coming from the scientific environment regarding the possibility of forecasting
and protecting the population, notification of the local and central authorities on the
possibility of natural disasters in order to prevent and reduce human and material damage and
to protect the environment. The difficulties encountered by local authorities are the following:
occurrence of unpredictable natural phenomena such as heavy rain, storms, tornadoes, floods,
earthquakes, extreme temperatures, abundant snow, climate change phenomena, etc ;
Insufficient funds to support rigorous research work in this area; lack of resources (equipment
and staff ); lack of expertise..
In order to realize the activity of setting up the expert system for sustainable
development, the exploitation of risk maps in order to prevent natural disasters (floods,
earthquakes, landslides, droughts, extreme meteorological phenomena, etc ...) in the pilot
area, all started from the construction of a graphic database including descriptive information
on the studied phenomena. These databases are correlated and operate within a GIS
application. Cartographic data are geospatial and alphanumeric data contain the characteristic
information for each type of natural risk that has been studied.
Geospatial cartographic data are collected using geodetic-topographic methods and
equipment photogrammetric and remote-sensing methods, and alpha-numeric data are
retrieved by purchasing them from statistics, from local government and directly from the
ground, by determining the level of groundwater, humidity through specific methods and
instruments (for example, in order to highlight the landslides, sensors are being placed in
locations established before, based on study cases, connected to a real-time monitoring system
and on its base when the favorable conditions for the occurrence of the phenomena show up,
the optimal decisions can be made. Alphanumeric data refers to three categories of
parameters: favorable to trigger natural disasters, to establish the vulnerability rate of the
predisposed areas, and for prevention
The results of the project are estimated to be used in the development of an expert
system for decision-making in real-time crisis situations, but also to prevent the occurrence of
human and material damage, imposing restrictive conditions on urbanism policies (location of
buildings, infrastructure development, etc.), and the establishment of specific measures to
reduce the effects.

2. MAKING A CARTOGRAPHIC EXPERT SYSTEM

An expert system is a program that uses knowledge and inferential procedures to

solve problems that are difficult enough that normally require the intervention of a human
expert to find the solution; in short, expert systems are programs that store specialized
knowledge introduced by experts

2.1. Implementation of the geodetic infrastructure

In order to achieve this infrastructure have been determined 4 points of support by

GNSS determinations, compatible to class C and 10 support points of class D (for cadastre- 6
buildable area and 4 in the built-up area).
 Fig.2. Integration of support networks into the national geodetic network

2.2. Updating the cadastral map

In order to update the topographical and cadastral map, we used the drone.

Fig.3. Workflow scheme for updating cartography using drones

Taking into account all of these we started the flight mission. To get the most accurate
results of the scanned reality, it is recommended to have a large number of aerial photographs,
because, in this way, when the image will be processed an accurate 3D model will be
obtained, and not an 2,5 D one, like in the majority of cases

2.3. Creating the cartographic data base as support for the expert system

Creating the information database of the integrated global resources, of the environment, on
a certain area of land both inside and outside the city, by realizing the fundamental
components of the informational system of the territory, namely:
• The general topographic map of the territory, including representation of relief.
.
 Fig.4. Topographic maps in the buildable pilot – area

• The cadastral map of real estate from buildable and built-up area ( project
application)
• Pedological map and pedological profiles

Fig.5. Pedological map and land profiles of the pilot-area

The main necessary product, used to achieve the goals set for this project is the digital
topographic map, which represents, by IT perspective the main graphic and descriptive
information database concerning the environment from the represented area on the map.
.

Fig..6. Derivative topographic maps of the objectives

 The digital map is made through a technological process specific to its content through
geospatial measurements, using modern equipment that is characteristic for the domains:
geodesy, topography, photogrammetry, remote sensing and mapping, and materializes by
classical processes on paper, plastic, etc. and / or modern electronic procedures.
The Digital Terrain Model (DTM) is a mathematical tool that allows numerical / digital
representation of any portion of the terrestrial surface. From the IT perspective, a numerical
terrain model appears as a file where data is
organized as a matrix in which each value
represents the altitude of a point. Starting from
this altimetric database, a large number of
products can be obtained, especially regarding:
slope, exhibition, perspectives, profiles through
different sections, etc.. In GIS, these models are a
fundamental layer because topographic
parameters determine a large number of
processes such as: leakage on the slopes, erosion,
sedimentation etc. permanent, geo - topographical
support points necessary for the topographic map.
The types of data contained in the graphical and non-graphic databases of the Expert
System for the development and exploitation of risk maps in order to prevent natural disasters
(floods, earthquakes, landslides, droughts, extreme meteorological phenomena, etc.) in the
pilot area are shown in the scheme below::

Fig. 7. Types of data contained in graphic and non-graphic databases of the Expert System

General structure of an E.S.

An expert system must contain three main modules, which form the so-called essential system:

⚫ Knowledge base
⚫ Inference engine
⚫ Facts base

2.4. The Knowledge Base

• The knowledge base is made up of all the specialized knowledge introduced by the
human expert
• The knowledge stored here are mainly the descriptions of objects and the relationships
between them
 • The knowledge base is part of the cognitive system, knowing being stored in a
specially organized space
The form of storage must ensure the search for the knowledge pieces specified directly
by identifying symbols or indirectly through associated properties or inferences that start
from other pieces of knowledge
For example, here is how to build the knowledge base to combat soil erosion:

The WEPP (Water Erosion Forecasting Program) was conducted by the US Agricultural
Research Service to estimate the risk of erosion of sloping land. Several versions have been
developed to improve computer algorithms. It is supported by major US agencies, tested in
many laboratories in the US, Australia and Europe. The WEPP program uses a complex
database (75 parameters) that collects climatic, relief, soil, plant, agro-technical, irrigation,
drainage and soil erosion records. The main input and output parameters of the WEPP and
EPIC models are the following (Table 1):
Table 1

INPUTS

1. Relief data: Area, Latitude and longitude of river basin; aspect, slope, slope length; length slope
hydrographic network
3. Pedological data: Number of horizons
2. Climate data: Extreme precipitation (EPIC, (EPIC, WEPP); Texture (EPIC; WEPP);
WEPP) Albedo (EPIC; WEPP); Erosion (EPIC;
Rainfall (EPIC; WEPP) WEPP); Apparent Density (EPIC; WEPP);
Average rainfall (WEPP) Field capacity (EPIC; WEPP); Organic
Torrential Core Intensity (WEPP) Matter Content (EPIC, WEPP); Content of
Maximum and minimum temperatures (EPIC; rock fragments (EPIC; WEPP); Cationic
WEPP) exchange capacity (EPIC; WEPP);
Dew Point Temperature (WEPP) Concentration of organic nitrogen, nitrates,
Wind speed and direction (EPIC; WEPP) mobile phosphorus, organic phosphorus,
organic carbon, calcium carbonate,
vegetable waste, etc. (EPIC)
5. Vineyard data: crop rotation (EPIC;
4. Type of works: (EPIC; WEPP); Number of WEPP); Soil Coefficient with Vineyard
jobs (EPIC, WEPP) Workout Date (EPIC; Plants (EPIC; WEPP); Cutting height at
WEPP); Depth of soil mobilization (EPIC; harvest (EPIC; WEPP); Plane dimensions -
WEPP); Number of combine elements, seed coefficient (EPIC; WEPP); Growth Ratio
drills, etc. (EPIC, WEPP); Irrigation (EPIC, (EPIC; WEPP); Plane spacing (EPIC;
WEPP); Drainage (EPIC; WEPP); Fertilization WEPP); Depth of penetration of roots
(EPIC); Pesticide Treatments (EPIC) (EPIC; WEPP); Frost Temperature (EPIC;
WEPP); Stress Production (EPIC, WEPP)
OUTPUTS
- Soil leakage (EPIC, WEPP);
- Erosion (maximum and average deviation, maximum and average deposition) (EPIC, WEPP);
- Wind erosion (EPIC);
- Sediment Production (WEPP);
Specific sedimentation weight (WEPP);
- Sediment granulometry (WEPP);
- evapotranspiration (EPIC);
- Nitrogen and phosphorus lost through liquid and solid leakage (EPIC);
- Plant-derived nitrogen and phosphorus (EPIC);
- Agricultural production (EPIC);
- Amount of biomass (EPIC);
- Soil and Soil Residues (EPIC), etc.
2.5. The Interference Engine

The inference engine (or mechanism) represents the novelty of the expert systems, it takes
from the knowledge base the facts used to construct the reasoning, the inference mechanism
follows a series of major objectives, such as:
⚫ choosing the control strategy according to the current problem
⚫ developing the plan to solve the problem as needed
⚫ switching from one control strategy to another
⚫ executing the actions provided in the resolution plan

Although the mechanism of inference consists of a set of procedures in the usual sense
of the term, the use of knowledge is not provided by the program but depends on the
knowledge it has at its disposal.
For example, we chose to set up the typical meteorological year for the pilot area required to
determine the flood curves. In order to take into account as many meteorological
characteristics a weight is made. Taking into account the algorithm given above, a
representative month of a given year is obtained.
These 12 months are concatenated and the leveling at the transition from one typical month to
another is done by a weighted mean of the neighboring values with the weights w = [0.3, 0.4,
0.3].
If n is the number of data involved in the leveling operation (for us n = 8, the last 4 days of
a month and the first 4 days of the next month, the calculation sequence is:
for i=2:n-1
date1(i)=w(1)*date(i-1)+w(2)*date(i)+w(3)*date(i+1);
end
where data is the vector of the initial values and data1 is the vector of the leveled values

The temperatures amplitudes

12

10

0
ian feb mar apr mai iun iul aug sept oct nov dec

atm 7.38 7.82 7.24 10.99 9.17 11.01 10.15 9.84 9.89 8.47 7.44 6.48
2006 5.5 6.5 7 8.2 8.7 9.2 9.2 9.8 8.7 8.7 7.5 6.5

2.5.1. The decision making model

The complexity of economic, biological, technological processes, the race for market
supremacy makes it impossible to solve problems through the art of leadership, with the help
of common intuition and judgment. All this needs to be solved in the conditions of increasing
structural and functional complexity of the systems, the raising of the level of technicality
and corresponding to some accentuated specializations of the professions. It becomes
necessary to develop the scientific decision on the basis of calculations made with measurable
and quantifiable sizes, with the purpose of choosing a variant from which an optimal result is
obtained.
Economic performance in agriculture depends, primarily, on the availability and
quality of natural resources. Environmental damage can affect in a negative form the
agricultural productivity and revenue, which can put pressure on the economic and social
structure in rural communities.
An important category of IT tools used to solve problems related to sustainable
development in agriculture are decision support systems. These tools are also called Decision
Support Systems - SSD (Decision Support Systems ).
The multi-criterial decision model consists in choosing the optimal variant of a finite
set of variants that compare to each other in relation to a finite set of criteria, each pursuing a
goal: the maximum or the minimum.
The decision-making process involves evaluating several decisional variants in order
to choose one. As an optimal variant in relation to a criteria can be suboptimal with respect to
other criteria, it is looking for the variant which matches the best compromise for all criteria.

Starting from the matrix of consequences (aij) with i = 1,2, ..., m and j = 1,2, ..., n
and how each criterion can be expressed in a variety of units of measurement, we need the
normalized matrix the consequences (qij) with (i, j)  (1,2, ..., n) x (1,2, ..., m) by following
transformations:
• for the criteria that follow the maximum the applied transformations are:
aij
qij = sum , unde aijsum =  aij (1)
aij j
- and for the minimum criteria:
(2)

It is noted that normalized with this procedure qij  (0, 1] and

We calculate the entropy of the normalized consequences for each criteria:

−1 n (3)
hj = . qij . ln(qij ), cu  h j  [0,1]
ln(m) i =1

The type of normalization chosen makes it possible to use formula (3) for this indicator of
information, being fulfilled then the condition h j  [0,1]
The degree of diversification of information given by the results of criteria j can be defined
by

d j =1 -hj (4)

dj
pj = n

d
j =1
j (5)

We obtain the coefficients of importance (p0j) where we took into consideration the weights
given by the decident λ = ( 𝝀1, λ2,…, λn ). The evaluation of the importance of the criteria is
done by means of the coefficients of significance j = 1,2,….. n There are situations in which
not all criteria are equally important and a certain degree of subjectivity is introduced by
the intervention of the decision-maker.
Tj .p j with n and n (6)
po j = n
 T
j .p j
 po j = 1   j = 1
j=1 j =1
j=1

The utility function is calculated by an additive counting method. The optimal variant is
the one that
n
corresponds to the maximum utility
ui =  po j .qij with i = 1,2, .., m. (7)
j =1

2.5.2. Data Mining, Cluster Analysis, Nonlinear Projection

We click the Multi-Criteria choice. The dialog that serves to input the input data has on the
right the brief presentation of the steps to be followed for entering, correcting, processing the
data, showing the utilities of each variant, keeping the data and the results in a document with
the .txt extension for storage or printing, resuming the algorithm for another processing.
Data Mining: a process of extracting new information from existing data collections,
information that can answer both questions: What is happening? but also why it's happening?

]
 Recent computer science achievements have been marked by a reorientation in the
use of accumulated data volumes, moving from retrospective research to research that
addresses future issues, allowed by the evolution of Data Mining technologies
 Data Mining classifies and aggregates data from different and even incompatible
systems, looking for new associations.
 Advanced analysis techniques are used including statistical algorithms, artificial
intelligence, neuronal networks, fuzzy logic, genetic algorithms, data visualization.

A special place among the tools with which Data Mining operates is Cluster Analysis

Cluster Analysis: Seeks to divide a heterogeneous population into more

homogeneous groups called "clusters" and represents "a set of clustering methods
based on the similarities of their characteristics, organizing them in as fewer
submultiples as possible.
This technique has the ability to reveal hidden features in a set of recordings, despite
the volume and variety of details.
It can have several advantages including providing classifications, formulating
conceptual schemes, generating hypotheses or testing them.
To divide the data in clusters there is a number of terms with similar meaning like
cluster analysis, automated sorting, numerical taxonomy (taxonomy is a science
dealing with classifying and systematizing laws of real-world domains with a
complex structure, taxology ), and typological analysis.
It is important to establish the purpose of cluster analysis, the selection of variables
that serve as grouping criteria, object similarity measures, grouping methods
(grouping algorithms), group counting, and group interpretation.
Nonlinear Projection: A way to visualize groups, if they exist, is to choose as many points in
the plan so that the mutual distances between points are as close as possible to the spacing
between the points in the dimensional space
  Practically, the points in the map are randomly chosen, in equal number to the
points in the dimensional space. Then, the points in the map change so that the
distances between them are spaced by the spacing between the corresponding points
in the dimensional space.
 This type of point representation of the points will be called nonlinear projection.
 Read the input data file. The data is in the form of a matrix where the characteristics
of an object are numbered on each line
 Let be the number of features equal to p and the number of objects equal to n.
 In principle, randomly choose n points in the map (as many as the number of objects
with given characteristics). Let y be the matrix with the coordinates of these points
in the map. In order to avoid that different runs have different results and to improve
the convergence of the algorithm we proposed the choice of the initial points in the
map as the linear projection of the points in the p-dimensional space on the map of
the first two factorial axes.
 Is determined the spacing between objects in the dimensional space of the
characteristics and forms a matrix of these distances.
 The sum of these distances is determined.
 The points in the map will change so that the distances between them are close to the
distances d (i, j) between the objects in the p-dimensional space.
 Finally, we obtain the planar configuration representing the nonlinear projection of
the initial points from the p-dimensional space.
.We have introduced the data analyzed with VIP1 (shown below) in the Cluster
Analysis program for the nonlinear projection.

We use the linear projection model to compare the 12 months of TMY achieved for the pilot
area. Comparing elements are: air temperature (mean, minimum, maximum), amplitude,
rainfall, duration of sunshine
The result: We notice three clusters (three possible seasons)
Cold one with November, December, January, February and March, warm with May, June,
July, August and the transition season with April, September, October. One month of
transition from cold to hot (April) and the transition from hot period to cold period is done in
September and October.

One month of transition from cold to hot (April) and the transition from hot period to cold
period is done in September and October.
  On a multicriteria analysis model defined on information from the knowledge base
achieved on IMEV project with the goal of choosing the optimal rootstock for the
Sauvignon variety, we tested the correct functioning of the two decision making
software created by the Mathematics group: VIP1 and the Nonlinear Projection,
applied here as a preprocessing data procedure as well as an alternative.

 The second Nonlinear Projection application signals another consequence of

climatic changes: no longer identifies the four traditional seasons.

2.6. Facts base

• The fact base is an auxiliary memory
that contains all the user's data (the
initial facts describing the statement
of the problem to be solved) and the
intermediate results produced during
the reasoning.
• Its content is
generally stored in
volatile memory
(RAM), but at the
user's request, it can
be preserved (on hard
disk).
In our case the history of natural disasters:
floods and landslides.

Fig.8. Database structure for historical floods

2.7. Database analysis and system implementation for decision making

Planning risks management activities is important for the level, type and visibility of risk
management to be proportionate to their risks and importance.
The result will be a risk management plan - that will address methodological issues (defining
approaches, tools and sources of information to be used in risk management processes),
responsibilities, budgeting, programming of interventions, etc.
. A risk management information system includes the following modules::
Entreprise Relationship Planning-ERP, which integrates one of the main processes in the field
of economics and finance: funding sources, material and human resource management,
planning and monitoring of activities, etc.
- Customer Relationship Management (CRM) - the entire set of interactions between
decision makers and operators). CRM facilitates the provision of services via the Internet,
telephone, ATM - Automated Teller Machine / Kiosk, etc .;
- E-Business - the web access of ERP systems. It can take one of the forms: direct links
between information from risk areas (B2B), historical information from existing and up-to-
date databases (B2C), knowledge bases about the risk factors (natural calamities - SCM) –
fig.9,
- Business Intelligence (BI) – applications for data collection, storage and processing
in order to make decisions. BI applications include activities such as:
Decision Suport Systems- SSD
Online Analytical Processing – OLAP
Data Warehouse
Data Mining
As it can be seen, there is a relationship between ERP, CRM, E-Business and
Business Intelligence components Business Intelligence.

Fig 9. The Relationship between ERP, CRM, E-Business and Business Intelligence components for
Hazard and Risk in Natural Disasters
CRM systems interact with ERP components to perform the desired functions. In this
respect (B2C or B2B), calls ERP`s for processing demands or ERP processed data are
periodically archived in warehouse repositories and processed using OLAP tools etc..
The results of analyzes are used by BI components to make decisions and develop
action plans to minimize the effects of the risks.

3.CONCLUZII

This project seeks to provide the decision makers the tools needed to achieve the
objectives of sustainable development of the area defined by the ATU by:
 • develops a formal information system for the conservation of the natural resources of
the habitat, the economic and social development, the protection of the environment
and the prevention of disasters;
• it integrates within the Rural Center for Sustainable Development and it is the
possibility of studying and informing both decision makers and citizens on the
possibilities for identifying and mitigating the effects of natural disasters and poverty
facing the studied area by knowing the potential for sustainable development of the
area represents a complex informational integrated system by creating a support-
digital map- in which the environmental information is represented spatially,
overlapping cadastral data, data on the underground with the specification
of the geomorphological structure and the geological resources, data concerning
the potential of the land by pedological soil mapping, ambient environment, data
about climate, hydrography, etc., which allows an impact analysis of
disruptive climate factors, soil pollution factors, water, air pollution, allows
the economic evaluation of resources in order to eradicate poverty, create
opportunities for development, create jobs and encouraging development
initiatives of the local economy, and correlated with social data and information, the
system can establish through a specific zoning the vulnerabilities to discrimination
concerning the equality of opportunities, access to education, health insurance, or
fight against poverty.
The main results obtained through this project are the following:
making the updated digital map of the area;
the implementation of the information system for monitoring climate change;
creating an interface between the computer system and users, allowing active participation in
the collection of meteorological data and information, their introduction nto the system and
processing them with specialized programs, in order to zone the study area to obtain the
following spatial information:
• flood risk areas
• areas with risk of landslide
• polluted zones
• areas affected by erosion
• deforested areas
• classification of land by use categories and soil quality classes, etc.
• identifying the real possibilities for removing the "disparities" by highlighting the
economic resources of the area
• NGO collaboration with representatives of the local administration to identify factors
of economic progress
Addressed to: Decision makers for the economic and social development of the area, all
citizens in the proposed area.The project will be a pilot project for generalized application in
other areas.
Bibliography

[1] Armeanu Ileana, Petrehus Viorel** Modele matematice utilizate in managementul

exploatatiilor viticole*Universitatea de Stiinte Agronomice si Medicina Veterinara-
Bucuresti, Catedra de matematica si informatica**Universitatea Tehnica de Constructii –
Bucuresti, Catedra de Matematica si Informatica
[2] Balteanu D. s.a., Sistem informaţional geografic (GIS) pentru studiul dezastrelor
naturale-Program de cercetare
[3] Barbieru Adrian s.a.Identificarea şi delimitarea hazardurilor naturale (cutremure
alunecări de teren şi inundaţii). Hărţi de hazard la nivelul teritoriului judeţean.Proiect
MTTC - IPTANA -CNGCFT
[4] Barbieru Adrian s.a.Identificarea şi delimitarea hazardurilor naturale (cutremure
alunecări de teren şi inundaţii). Hărţi de hazard la nivelul teritoriului judeţean.Proiect
MTTC - IPTANA –CNGCFT
[5] Borcean I., Tabără V., David Gh., Borcean Eugenia, Ţărău., 1996, Zonarea, cultivarea şi
protecţia plantelor de câmp în Banat, Ed. Mirton, Timişoara
[6] Curcubăt Stela,2015, Evoluţia interpretării cartografice a învelişului de sol şi utilizării
practice a hărţilor pedologice,Teză de doctorat,
[7] Lelarin Julien, Hydrodynamic simulation with MIKE 21 –SOPAC trainining report87-
august 2000
[8] Murariu G, Praisler M,Stoian I, Numerical Simulation for Charges Flow Assessment
Using Fluent Platform, The Annals of the University Dunărea de Jos of Galaţi,
Mathematics, Physics, Chemistry, Informatics, Fascicle II, Supplement, Year III 2009,
Proceedings of the 1st International Symposium on Applied Physics Materials Science,
Environment and Health (ISAP1), Galati, November 28th-29th, 2009, p.121-125
[9] Stoian I: Realizarea hărţilor tematice, a hărţilor de risc şi managementul crizelor utilizând
tehnici de teledetecţie”, Revista de Geodezie, Cartografie şi Cadastru, 17-18 nov. 2005;
[10] Stoian I ”Elaborarea hărţilor de hazard la nivel judeţean”, Buletinul de Fotogrammetrie şi
teledetecţie nr. 32, Bucureşti, 2006
[11] Stoian, I. (2008): “Sistem suport de decizii pe baza hărţilor de hazard şi risc pentru
diminuarea efectelor calamităţilor naturale” –ANALELE CNGFT
[12] Stoian, I. M. Praisler, A. Anghelescu, V. Nacu, N. Baroiu- Monitorizarea hazardurilor şi
riscurilor în caz de calamităţi naturale în arealul Dunărea deJos-Proiect de cercetare
RISKGAL
[13]Xihua Yang and Bengt Rystedt: Predicting Flood Inundation and Risk Using GIS and
Hydrodynamic
Model: A Case Study at Eskilstuna, Sweden- Department of Technology, University of Gävle,
SE-801 76 Gävle, Sweden
([14] WCED)- 1983- raportul „Viitorul nostru comun”- Raportul Brundtland

http://www.scritub.com/tehnica-mecanica/SISTEME-EXPERT-SI-NEUROEXPERT441911164.php
http://www.utgjiu.ro/math/mbuneci/book/exp/cap08.pdf
http://www.mpt.upt.ro/doc/curs/gp/Sisteme_inteligente_in_electrotehnica/Inteligenta_artificiala_si_Retele
_neuronale_cap1.pdf
http://www.matrixrom.ro/romanian/editura/domenii/cuprins.php?cuprins=SF90
http://steconomiceuoradea.ro/anale/volume/2005/statistica-si-informatica-economica/13.pdf
http://safetyinknowledge.blogspot.ro/2015/01/dezvoltarea-unui-sistem-expert-pentru.html
http://www.revista.forter.ro/2009_3_t/06-arm/01.htm
https://www.icpa.ro/proiecte/Proiecte%20nationale/ExET/ExET14.pdf