TOWARDS A REPRESENTATION OF ENVIROMENTAL

MODELS USING SPECIFICATION AND DESCRIPTION

LANGUAGE

From the Fibonacci Model to a Wildfire Model

Pau Fonseca i Casas, Màxim Colls and Josep Casanovas

Universitat Politècnica de Catalunya, Jordi Girona 1, 3, Barcelona, Catalunya, Spain

Keywords: SDL, Simulation, Formalisms, Cellular automata.

Abstract: In this paper we explore how we can use Specification and Description Language (SDL) to represent

environmental models. Since the main concern in this kind of models is the representation of the

geographical information data, we analyze how we can represent this information in the SDL diagrams. We

base our approach using two examples, a representation of the Fibonacci function using a cellular

automaton, and the representation of a wildfire model. To achieve this we propose the use of a language

extension to Specification and Description Language. This allows the simplification of the representation of

cellular automatons. Thanks this we can define the behavior of environmental models in a graphical way

allowing its complete and unambiguous description. SDL is a modern object oriented formalism that allows

the definition of distributed systems. It has focused on the modeling of reactive, state/event driven systems,

and has been standardized by the International Telecommunications Union (ITU) in the Z.100.

1 INTRODUCTION

From the different phases of a simulation model

construction, the formalization phase, sometimes is

missed. This phase is needed in order to understand

the model before any implementation. The

specification helps in the implementation process

and in the communication between the different

personnel involved in the model construction. Also

can be considered a product itself (Brade, 2000).

Different formalisms exist in order to represent a

simulation model, like Petri Nets or DEVS among

others. Some tools have been built in order to allows

help the model implementation from the

specification (De Lara, et al., 2002), (Praehofer, et

al., 1993) and some allows the distribute execution

of the models like CD++ (Wainer, et al., 2003).

We select Specification and Description Language

in order to represent our simulation, an ISO

language than allow representing the models

graphically. SDL (Doldi, 2003),

(Telecommunication standardization sector of ITU,

1999) is the acronym of Specification and

Description Language. SDL is an object-oriented,

formal language defined by the International

Telecommunication Union – Telecommunication

Standardization Sector (ITU–T) (formerly Comité

Consultatif International Télégraphique et

Téléphonique [CCITT]) as Recommendation Z.100

(Telecommunication standardization sector of ITU,

1999). The language is designed to specify complex,

event-driven, real-time, interactive applications

involving many concurrent activities using discrete

signals to enable communication.

SDL is a powerful and modern language widely

used in different areas, not only in simulation area. It

has been standardized by the International

Telecommunications Union (ITU) in the Z.100, and

can be used easily in combination with UML.

The definition of the model is based on different

components:

• Structure: system, blocks, processes and

processes hierarchy.

• Behavior: defined through the different

processes.

• Data: based on Abstract Data Types (ADT).

• Communication: signals, with the parameters

and channels that the signals use to travel.

343

Fonseca i Casas P., Colls M. and Casanovas J..

TOWARDS A REPRESENTATION OF ENVIROMENTAL MODELS USING SPECIFICATION AND DESCRIPTION LANGUAGE - From the Fibonacci

Model to a Wildﬁre Model.

DOI: 10.5220/0003066203430346

In Proceedings of the International Conference on Knowledge Engineering and Ontology Development (KEOD-2010), pages 343-346

ISBN: 978-989-8425-29-4

 2010 SCITEPRESS (Science and Technology Publications, Lda.)

• Inheritances: to describe the relationships

between, and specialization of, the model

elements.

The language has 4 levels (i) System, (ii) Blocks,

(iii) Processes and (iv) Procedures (Figure 1).

Figure 1: SDL levels (IEC)

To know more about SDL the recommendation

Z.100 (Telecommunication standardization sector of

ITU, 1999) can be consulted, also a lot of

information can be reviewed in the www.sdl-

forum.org website.

In our domain the signals are equivalent to the

events that rule the modification of the states of the

different elements, for that in this paper the use of

event or signal is equivalent.

1.1 Cellular Automata

Cellular automatons must be defined in SDL

because simplify the interaction of simulation

models and GIS data (Benenson, et al., 2004). The

cellular automaton we are using is an extension of

the common cellular automaton named m:n-CA

. Its

definition can be found on (Fonseca i Casas, et al.,

2005). Cellular automata are discrete dynamical

systems whose behavior is completely specified in

terms of a local relation (Emmeche, 1998).

One-dimensional cellular automata are based in a

row of "cells" and a set of "rules". A two-

dimensional cellular automata uses rectangular grids

of cells. Each one of the different cells can be in one

of different "states" (the number of possible states

depends on the automata). Thinking states as

numbers, in a two-state automata, each cell can be

only in 1 or 2 state. Cells represent automata space;

time advances in discrete steps following “the

rules”, the laws of “automata universe”, usually

expressed in a small look-up table. At each step

every cell computes its new state in function of its

closer neighbors. Thus, system's laws are local and

uniform.

2 IMPLEMENTATION

To implement our models we can use different

existing tools that understand SDL language, like

Cinderella (CINDERELLA SOFTWARE, 2007) or

Telelogic (IBM, 2009) of IBM. We develop our tool

in order to improve the existing solutions adding

some new capabilities:

• Allow to work with the delaying signals (as we

can see below).

• Allow to work with cellular automata.

• Allow to work with intelligent agents.

• Allow a distributed simulation of the models.

For these reasons we decide to implement our tool

named SDLPS (Fonseca i Casas, 2008).

2.1 Proposed SDL Extensions

In SDLPS all the signals can carry the parameter

defined in the structure represented in the Figure 2.

Figure 2: Structure related to the SDLPS signals.

ExecutionTime, of event structure, allows defining

the time when the signal (carrying the event) must

reach its destination. Other elements are: (ii)

Priority, the priority of the event, used to break a

possible simultaneity of events. (iii) CreationTime,

representing the time when the event is created. (iv)

Id, an identifier of the event. (v) Time, the clock of

the process. (vi) Destination, the final destination

(process PId) of the signal. Not all the parameters of

event structure must be defined, only those needed

to fully define the behavior of the model.

Figure 3: Defining the executionTime, and other

parameters, of the signal using the SDL time extensions.

Regarding the implementation for the cellular

automaton remark that all the cells have the same

behavior. This implies that is not needed to represent

all the cells, but only one cell. Also is needed to

represent the relation that have with the



KEOD 2010 - International Conference on Knowledge Engineering and Ontology Development

344

neighborhood (that of course is specific of the

cellular automata), and the relation with the other

layers. That means define the combination, vicinity

and nucleus functions (Fonseca i Casas, et al.,

2005).

In SDL language we can use types to define blocs

that have the same behavior (as is usual in any OO

language).In order to represent the communication

between the different cells, are needed to represent

all the communication channels between the cells.

This implies the need of represent at least a set of

cells that belongs to the vicinity (and the nucleus) of

a cell. Also if we want a complete description of the

cellular automaton (if we want an automatic

generation of code), we need to represent all the

cells of the cellular automaton.

To simplify the representation (avoiding the

representation of all the cellular automaton cells) we

decide to add a new kind of agent to SDL language,

the mnca. This agent have the same behavior as the

agent block, with a particularity, mnca agent is

defining the entire cellular automaton, but only is

needed to represent one cell. This implies that is

needed to use a declarations section that defines the

structure of the cellular automata (mainly the

dimension of the cellular automata, and the size of

each one of these dimensions). Also the OUTPUT

SDL elements can carry a parameter to define what

is the cell that receives the signal (Fonseca i Casas,

et al., 2010). We can define an array of cells (the

cells that can receive the signal). Also we can use

ALL_CELLS, to send the signal to all the cells of

the mnca agent.

3 FIBONACCI FUNCION MODEL

As a first example we represent cellular automata

that calculate the well known Fibonacci function.

First is needed to define the different elements that

compose our simulation model.

This first level of the SDL diagrams, in this case,

only contains a single block, representing the

cellular automata that implements the Fibonacci

function. Always this first level of the model helps

to understand the different elements that must be

combined inside the model and the interaction of the

model with the environment (users, computers, etc).

Next, we must define the structure of this m:n-CA

cellular automata. First the number of cells (the

dimensions), using the DCL (declaration block). On

this block, the mnca_DIM variable defines the

number of dimensions of the cellular automata, and

mnca_D1 to mnca_Dn defines the size of each one

Figure 4: m:n-CAk cell representation. Here we can see

the relation between the different layers.

of these dimensions. In that case we have a matrix

(10x10) as is represented in Figure 4.

Evolution function is defined in the ProcessLayer, to

see its representation (Fonseca i Casas, et al., 2010)

can be consulted.

4 WILDFIRE MODEL

As a second example we show the first diagrams of

a wildfire model. In that case we are following

behave model to represent the fire spread (Andrews,

et al., 1989). Is interesting to remark that since in

this model we need different layers we must define

them on the DCL block on the block mnca (Figure

5).

Figure 5: m:n-CAk cell representation for wildfire model.

Inside this we can find the definition of the behavior

of the model. For space reasons we only show the

states diagram that define the transition on each one

of the cellular automaton cells (Figure 6). Also, on

Figure 7 we show the aspect of the SEND signal that

represents the propagation of the fire to other cells of

the cellular automaton.

TOWARDS A REPRESENTATION OF ENVIROMENTAL MODELS USING SPECIFICATION AND DESCRIPTION

LANGUAGE

345

Figure 6: states diagram for a cell.

Figure 7: propagation of the wildfire.

5 CONCLUSIONS AND FUTURE

WORK

This paper shows how we can model environmental

systems using Specification and Description

Language. To do this the main concern is how

model the behavior of cellular automata graphically

using SDL, and how to manage time.

Two examples are quoted, a representation of a

Fibonacci function over a cellular automaton, and

the fire spread following the Behave model.

Two proposed extensions to SDL are shown, one to

manage time on the SDL signals and other to

simplify the representation of the cellular automaton

structures.

The future work is focused in the verification of the

implemented structures on SDLPS and the use of

this system on some ongoing real projects involving

industrial models or environmental models.

REFERENCES

Andrews P. L. and Chase C. H. Behave: Fire behavior

prediction and fuel modeling system-BURN

subsystem, part 2 (Report): Gen. Tech. Rep. INT-260..

- Ogden, UT: U.S. Department of Agriculture, Forest

Service, Intermountain Research Station, 1989. - p. 93.

Benenson Itzhak and Torrens Paul M. Geosimulation,

Automata-based Modeling of Urban Phenomena

(Book) - West Susses PO19 8SQ: John Wiley & Sons

Ltd., 2004. - ISBN: 0-470-84349-7.

Brade D. Enhancing modeling and simulation

accreditation by structuring verification and validation

results (Conference), Winter Simulation Conference /

ed. Joines J. A. (et al.). - 2000.

CINDERELLA SOFTWARE Cinderella SDL (Online). -

2007. - 03 31, 2009 - http://www.cinderella.dk.

De Lara J. and Vangheluwe H. ATOM³: A Tool for Multi-

formalism Modelling and Meta-modelling

(Conference), 4th International Conference on

Enterprise Information Systems ICEIS 2002. - 2002.

Doldi Lauren Validation of Communications Systems with

SDL: The Art of SDL Simulation and Reachability

Analysis (Book) - [s.l.]: John Wiley & Sons, Inc.,

2003 - p. 310. - ISBN: 978-0-470-85286-6.

Emmeche Claus Vida Simulada en el ordenador (Book). -

Barcelona : Gedisa, 1998. - ISBN: 84-7432-616-8.

Fonseca i Casas Pau and Casanovas Josep Simplifying

GIS data use inside discrete event simulation model

through m_n-AC cellular automaton (Conference) //

Proceedings ESS 2005. - 2005.

Fonseca i Casas Pau SDL distributed simulator

[Conference], Winter Simulation Conference 2008. -

Miami: INFORMS, 2008. - http://wintersim.org/

abstracts08/POS.htm#fonsecaicasasp84590.

Fonseca i Casas Pau, Colls Màxim and Casanovas Josep

Representing Fibonacci function through cellular

automata using Specification and Description

Language (Conference), Procedings of the 2010

Summer Simulation Multiconference. - Ottawa, ON,

Canada : (s.n.), 2010.

IBM TELELOGIC (Online). - 2009. - 03 31, 2009. - http:

//www.telelogic.com/

IEC SDL Tutorial (Online), IEC. - May 2010. -

http://www.iec.org/online/tutorials/sdl/topic04.html.

Praehofer H. and Pree D. Visual modeling of DEVS-based

multiformalism systems based on higraphs

(Conference), Winter Simulation Conference / ed.

Evans G. W. [et al.]. - Los Angeles, California, United

States: (s.n.), 1993. - pp. 595-603. - DOI= http://doi.

acm.org/10.1145/256563.256737.

Telecommunication standardization sector of ITU

Specification and Description Language (SDL)

(Online) = ITU-T Z.100. 1999, Series Z: Languages

and general software aspects for telecommunication

systems.. - International Telecommunication Union,

1999. - April 2008 - http://www.itu.int/ITU-

T/studygroups/com17/languages/index.html.

Wainer G. and Chen W. A framework for remote

execution and visualization of Cell-DEVS models

(Journal), Simulation: Transactions of the Society for

Modeling and Simulation International. - November

2003. - pp. 626-647.

KEOD 2010 - International Conference on Knowledge Engineering and Ontology Development

346