=Paper=
{{Paper
|id=Vol-52/paper-7
|storemode=property
|title=Ontological modelling of Natural Categories:an Ant Colony
|pdfUrl=https://ceur-ws.org/Vol-52/oas01-gutierrezcasorran.pdf
|volume=Vol-52
}}
==Ontological modelling of Natural Categories:an Ant Colony==
https://ceur-ws.org/Vol-52/oas01-gutierrezcasorran.pdf
Ontological modelling of Natural
Categories-based Agents: an Ant Colony
Cesáreo Gutiérrez-Casorrán Jesualdo Tomás Rodrigo Martínez-Béjar
Fernández-Breis
Departamento de Ingeniería de la Departamento de Ingeniería de la Departamento de Ingeniería de la
Información y las Comunicaciones, Información y las Comunicaciones, Información y las Comunicaciones,
Facultad de Informática, Campus de Facultad de Informática, Campus de Facultad de Informática, Campus de
Espinardo, Universidad de Murcia, CP Espinardo, Universidad de Murcia, CP Espinardo, Universidad de Murcia, CP
30071 Espinardo, Murcia, Spain 30071 Espinardo, Murcia, Spain 30071 Espinardo, Murcia, Spain
+34 968 367345 +34 968 364634
zesareo@retemail.es jfernand@perseo.dif.um.es rodrigo@dif.um.es
A more modern phenomenon in the representational theory field
ABSTRACT can be found in the usage of ontologies for sustaining the agent's
Nowadays, the study of multiagent systems is a significant and
knowledge [4]. The term ontology has a philosophical origin
promising research area, since they are considered useful for
since it comes from the Aristotle's attempt to make a
simulating the behaviour of real systems. On the other hand,
classification of the things in the world. In the AI field, many
ontologies are considered the best manner for representing
definitions have been given to the term ontology, although the
reusable and shareable knowledge. In this work, we make use of
most accepted one is that given in [14], where it is defined as "a
ontologies for modelling the multiagent system. In particular,
formal, explicit specification of a shared conceptualisation".
top-level ontological categories are used for modelling the
Moreover, ontologies provide potential terms for describing our
concepts of the domain. The application domain is an ant
knowledge about the domain [4].
colony, and some results of this research are shown in the paper.
The main advantage of using ontologies for representing the
Keywords existing knowledge is that they allow for sharing and reuse of
Multiagent Systems, Ontologies, Social Systems bodies of knowledge in a computational form [6]. The
knowledge included in an ontology can be shared because this
1. Introduction knowledge uses to be consensual, that is, it has been accepted by
Nowadays, both Agent and Multiagent Systems (MAS) fields a group not by a single individual [24]. Ontologies have also
constitute an important research and development area in an been used for building knowledge bases in different fields [12].
increasing number of Universities, Research Institutes and, from In the ontological engineering research area there are two main
the last decade, in the industria. In consequence, these ideas are approaches for building ontologies:
applicable into multiple domains. Furthermore, many definitions (a) Collaborative ontological development: Due to the
of what an agent and an MAS are, each being valid for a certain significance of the ontologies for supporting the representation
aplication domain but no one better or more correct than the of knowledge in CSCW tools, the importance of collaborative
others, have been given [19]. Moreover, MAS are considered a ontological engineering is increasing as it can be observed from
positive manner for modelling and simulating real systems [10]. the number of such a class of tools that have appeared lately
In this work, the concepts of agent and MAS are going to be ([6]; [8]). These systems are domain ontologies-oriented.
applied in the domain of artificial societies' systems that
simulate the behaviour of a specific type of natural society (i.e., b) Generic ontologies construction: Some attempts for creating
ants). large scale, generic ontologies that fix a series of concepts
according to some general classification criteria have been
Modelling agents for simulating artificial social systems has made. From this first classification of concepts, ontologies for
frequently been performed by means of simple agents, and some specific domains can be developed. The advantage of this type
interesting results have been drawn from it [7]. However, such of ontologies radixes in the resolution of compatibility problems
approaches impose strong constraints to the emergent social that appear when trying to unify ontologies designed according
behaviour, reducing the possibilities of analysing and obtaining to different high-level concepts. The project Cyc's goal [5] was
conclusions related to more complex societies as the human one to create an ontology of this type and used three classification
is. A similar fact happens with the traditional attempts made by basic criteria:
the Artificial Intelligence (AI) community in order to
incorporate rational behaviour to the agents by applying (1) Represented things versus internal things.
techniques such as logical rule sets or representational theory (2) Individuals versus collections.
inside the MAS architecture. Actually, it has usually been (3) Tangible versus intengible versus both.
necessary to simplify the problem due to the complexity
provoked by the application of such techniques from a In [20], some problems with the CYC ontology are raised: (1)
computational point of view [1; 2]. the category Thing has no properties of its own; (2) things
cannot be represented on its own machine; (3) collections are
treated as intangible; (4) Process is under Individual Object but
Tangible Object is under Process, etc. Thus, the top-level
ontology designed by Sowa attempts to overcome those
problems [21]. This particular ontology has been used as a
reference for our work.
As a counterpoint to the current state of the research and
development in the area, some metamodels have appeared with
the aim of solving the problems related to the traditional
Artificial Intelligence approaches. The author in [13] defines an
agent, but the novelty of this definition is that it gives reasons
that support the unnecessity of keeping a partial representation
of the environment. In [16], the author states that the
intelligence could emerge from the interaction between agents
with differentiated functionalities. More recently, Goldspink
addresses topics such as the incorporation of the complex
systems theory and the autopoiesis to solve those problems,
emphasizing in this way his job's intentionality, which has
become a new research direction to be used as the starting point Figure 1: Lattice of top-level categories
for further developments that concretizes the metamodel and
that can prove its utility for studying social systems. The Sowa's top-level categories [21] have been taken as a
reference for designing the agent's ontology. This author
The application domain chosen is an ant colony which is an defines three classification criteria for formalising the ontology:
application domain in which the indivudual level is not the most
important but the group level, because the ants together form a (a) It distinguishes among abstract and physical concepts
coherent whole and maintain themselves as a group whereas according to the spatial-temporal location. For example, an
individuals have no or little self-interest [23]. Then, the MAS abstract entity is the proper information, independently from the
modelled in this work corresponds to an ant colony. physical medium it is captured from. The same information can
be included in a tape, a newspaper, etc, although it is always the
The structure of this paper is as follows. The methodology used same information independently from its location in time and
for developing this work is described in Section 2. Section 3 place. On the other hand, a physical entity is defined as having
describes the coordination process in multiagent systems. In both a spatial and a temporal location. Other attempts for
Section 4, the scenario in which the model is applied is defining physical entities, such as the energy or the mass would
described. Section 5 reflects the results obtained by applying the imply the exhasutive study of physical theories.
methodology in a concrete application domain. Finally, some
conclusions and final remarks are put forward in Section 6. (b) According to the dependence degree with respect to other
entities, an entity can be either independent, relative, or
2. Methodology mediator. The existence of an independent entity does not
Given the current state of the art in the ontological agents' depend on being related to other entities. However, a relative
modelling it seems logic that if we maintain the classical trends entity can only exist if it is related to other entities. Finally, a
in the development of new projects, we should find the same mediator entity establishes a relationship between two entitites.
computational complexity problems. The relative novelty in the (c) According to the temporal stability of its identity, an entity
development of ontological agents is a good reason to continue can be either continuous or occurrent. An entity is continuous
to exploit techniques for simulating natural social systems by when its identity is recognizable during a non-empty temporal
means of artificial societies. The purpose then is to obtain interval. An entity is occurrent when it cannot be associated to a
conclusions about their benefits and to reach social systems concrete temporal interval (e.g., the human being’s life).
closer and closer to human systems, which is the most important
goal of this discipline. If we combine these distinctions, 12 concepts are formed,
namely: Object, Process, Schema, Script, Juncture,
Our work consisted in creating an artificial social system in an Participation, Description, History, Structure, Situation, Reason,
ants colony context. In this section, the following aspects are and Purpose. Figure 1 shows the lattice defined by the previous
presented: (1) the ontology; (2) the definition of agent; (3) the distinctions and the concepts derived from them.
social laws for coordinating the MAS; and (4) the application
scenario, the colony of Nothomyrmecia macrops ants. 2.2 A natural categories-based agent
We use in this work the definition of agent given in [21]: an
2.1 The top-level ontology agent is an animate entity capable of performing tasks with a
A model is ontological when it explicitly makes use of specific purpose.
ontologies. In principle, the ontology is the key element for
allowing the agents defined with this model to interpret the An animated entity is one that has soul or anima (from the Greek
incoming information, so that the agents can perform their tasks word psyche). Aristotle defined the psyche as the principle a
(e.g., planning, goals identification, etc) once such (incoming) living entity is determined by, but in his efforts for classifying
information is represented in a useful way. all the existing things, he also defined the processes an animated
entity should meet, namely: nutrition, perception, desire,
locomotion, imaginery and thought. In this work, the agents of
the artificial society will be assumed to possess the capabilities relevant aspects. Moreover, the agent will elaborate a (more or
described in the following lines. less general) plan to achieve each particular objective, or the
subset of objectives chosen.
2.2.1 Nutrition
Nutrition could be defined as the "Purpose"1 objective 2.3 The agent's control
established for the agent with the intention of keep maintaining The following types of control modules have been designed for
or improving their existence. This implies that the agent must the agent:
have knowledge of its own existence and be aware of the fact
that it is independent, actual, and continuous, that is, an object 2.3.1 Knowledge acquisition
"Object" according to Sowa’s top-level categories. Furthermore, The sensors of the agent receive information from the
it must be capable of measuring how good it feels in order to environment and compare this information with the current
know its capability to survive. agent’s state to identify the significant changes produced in the
environment. That new information is interpreted by making use
It is said that an agent is capable of performing the nutrition
of the ontology and the (new) concepts underlying the current
process when:
situation.
1) it knows about its own existence; and
The agent has a memory where the most relevant situations and
2) it knows its health condition; and concepts are stored. The current situation progressively forgets
3) it has defined objetives in order to be always in his best concepts and these concepts are included into the agent’s
health condition. memory. These concepts will gradually loose importance if they
are not reinforced and they will be finally forgotten.
2.2.2 Perception
Perception consists in receiving information from the 2.3.2 Monitoring for identifying objectives
environment. The agent must know its environment as an entity The agent evaluates the current situation and new objectives are
separated from itself, and it must periodically receive generated to improve that situation. Moreover, these objectives
information through its input channels "Juncture". stand as a part of the current situation. This module can also
discard current objectives or modify some parameters of such
It is said that an agent is capable of performing the perception objectives.
task when:
1) it knows about the existence of its environment; and 2.3.3 Planning and control
Plans are developed attending to the strategies defined for each
2) it has a set of input channels that connect the environment to
objective. Then, the decision of which action or actions to
the agent; and
perform at each moment is made. Each new situation implies the
3) it is capable of cataloguing the information from the revision of the plan and deciding again the action to perform.
environment in one of the concepts defined in the agent's
ontology. 2.4 The ontological framework
An existing ontological tool, called ONTOIN, has been used for
2.2.3 Desire the construction of the ontologies presented in this work because
By its definition, an agent wants to meet its objectives in order this tool allows for editing inconsistency-free ontologies in a
to reach a specific state, which has been considered to be better graphical, friendly manner [8;9]. The ontological schema
than the current one. The agent's state can be understood as its included in this tool can be described as follows. In ONTOIN,
current situation "Situation", which would be comprised of all an ontology is comprised of a set of concepts, each having a set
the instances interpreted in the last x time stamp units. of attributes. Each attribute has a range of possible values. An
It is said that an agent is capable of performing the desire attribute can be either specific to that concept or inherited from
process when it has defined a series of actions to decide which another concept of the same ontology. Here is where the
states are better than others. relationships appear in this ontological model, and a relationship
always involves the participation of two concepts.
2.2.4 Locomotion So far, ONTOIN allows for defining taxonomic (is a class of)
Locomotion is defined as the capability an entity has to move in and mereological (is a part of) relations, although this tool is
its environment in both temporal and spatial dimensions. It is currently being expanded to cover a wider range of relations,
said that an agent is capable of performing the locomotion such as topology, causality, influence, etc. Returning to the
process when: nature of the attributes, if an attribute A from a concept C1 is
1) it knows that it has determined spatial and temporal locations; inherited from a different concept C2 then there exists a
and taxonomic relation between C1 and C2, written IS-A (C1, C2).
2) it is capable of modifying its location voluntarilly. Moreover, in this model attribute (multiple) inheritance is
permitted for taxonomic organisations.
2.2.5 Thought On the one hand, taxonomic relations are assumed to hold all
Thought consists in identifying the relevant aspects of the irreflexivity, asymmetry and transitivity. On the other hand,
current situation and deducing the goals to achieve from those mereological relations are assumed to hold irreflexivity,
asymmetry and non-transitivity. The axiomatic ontological
1 component is also taken into account in ONTOIN since different
See [21] for further information about Top-Level Categories types of axioms can be represented in this ontological model: (1)
Internal structure-based axioms, that is, axioms that result from 3.2 Simple social law
the relation ‘concept has attribute’; and (2) other axioms, The simplicity criterion for social laws consists in creating
derived from some properties concerning relationships between useful social laws that allow the agents to make their own
concepts for taxonomic and mereological organisations. decisions by taking the least possible number of information
The ONTOIN tool does not only is useful for editing ontologies captured from the environment into account.
but it can also be used for integrating different ontologies A social law imposes a series of contraints over the actions an
specified with the edition facility. The integration framework agent must take under determined circumstances. A useful social
used in the ONTOIN tool is adapted from the one presented in law is said to be simpler than another useful social law when it
[9]. For this work, only the editing facility has been used since is easier for the agent to know whether an action can be done
our goal was to build the ontologies that represent the different than with the other law. By using simple social laws, the agents
agents involved in the MAS. are less dependent on their sensors and allow the agents to
develop simpler strategies.
3. Coordination in the MultiAgent System
A basic aspect in the development of a MAS is to coordinate all 3.3 The formalization of the system
the agents. The principle of rationality can be described as The following definitions establish the formal framework for the
follows: "if an agent has knowledge that one of its actions will concepts of the study of the agent's model used in this work.
lead to one of its goals, then the agent will select the action"
[17]. In order to apply it to an MAS, it must be redefined in 3.3.1 The environment
order to refer to the overall system. This is described by the The environment can be defined as a set of tuples with a specific topological structure. This structure is
immediately above the knowledge level, which is concerned determined by the connectivity of the cells. Let us now
with the inherently social aspects of the MAS" [15]. formalize the concepts location and cell:
There exist different approaches for modelling the social aspect Location: It is a bidimensional vector (x,y) that defines the cell
of the agents. For instance, in [3] the author models explicitly position in the environment. Two cells are said to be connected
social actions and in [18] new categories of social agents are (i.e, adjacents), when the arithmetic difference between any
presented. component of their location vectors is equal to 1.
The design of social laws for artificial social systems is a good Cell: It is a tuple . Attributes are the
approximation between a totally centralized approach and a physical properties of the environment, such as temperature,
totally distributed one. pressure, etc. On the other hand, the objects are the independent
In an artificial social system, each agent must decide according entities placed in that particular cell.
to its current state and to the state of its accessible environment.
In a distributed approach whatever conflict between agents
3.3.2 Objectives, strategies, and actions
The objective of an agent is a tuple
would imply to establish a negotiation process between the
where:
involved agents. Social laws impose constraints to the agents'
activities so that conflicts are avoided and they ensure that the Expression: It is a function f: S→Z, where S stands the agent's
agents are capable of achieving their objectives and the society states and Z is the set of positive and negative numbers. Thus,
can meet its global objective. given the current situation s ∈ S, f(s) is the completion degree of
Designing social laws is a quite complex task, so it would be the objective. The objective intends to reach a state in which f(s)
very helpful to have a metric for measuring the goodness of a is maximized.
particular social law in comparison to other laws. In [11], Strategy: It can be defined recursively as follows:
several criteria that have been considered useful are defined,
namely: minimal social laws criteria, and simple social laws
A function g: E x S → P(A), where E stands for the
environment state, S stands for the agent state, and
criteria.
P(A) stands for the power set of actions that can be
3.1 Minimal social law performed.
Each agent has both a series of objectives to achieve and a series A sequence of goals.
of strategies to accomplish those objectives. Besides, objectives An action defined on the environment (or on the agent) is a
at a social level have been defined, and these objectives must be function h: E→E (or h': A → P(A)), where E stands for the
met at every moment in order to ensure the survival of the environment state (converserly A stands for the agent states).
society. Therefore, an agent can perform actions either on the
A social law is useful when it guarantees the achievement of the environment or on other agents.
objectives at a social level and allows each individual agent to It can be said that there does not exist defined strategies for
achieve their individual goals in one manner at least. achieving social goals because it is the set of all the agents'
Thus, a social law is minimal when it is useful and it permits the strategies, which allows for the achievement of the expression.
agents to be as free as possible, that is, there is not another
useful social law that offers the agents more alternatives to 3.3.3 The social laws
achieve their objectives. In order to establish the social laws, an initial hypothesis must
be formulated. Our initial hypothesis is this: "with the defined
strategies, the social goals can be achieved. Moreover, each
individual agent can also achieve its own goals with those 3.3.4 The multiagent system
strategies". The multiagent system is a tuple ,
A social law is a subset of strategies which cannot be modified where:
by any agent individually. A social law is useful while the initial • E stands for the sequence of tuples
hypothesis is valid. A social law is minimal if there is not a
smaller subset of strategies that conforms a useful social law. A
social law is simple when the strategies that are permitted for the • Li: It is the sequence of tuples
agent uses the least possible information for establishing the • IC represents the set of initial conditions for both
possible actions to be performed. environment and the set of agents.
Furthermore, the social laws are incorporated into the • A is the set of actions that can be performed by an
ontological model as axioms of the domain, that is, they must agent on the environment and on other agents.
always hold in order to ensure the consistency and correct
evolution of the MAS. This guarantees that the social laws are • S stands for the set of strategies defined for the
held and the agents' interactions are constrained by these social system.
laws or axioms of their ontological model. • Oi represents the set of goals defined for each agent.
• Os is the set of social goals.
Figure 2. The natural categories for some of the most significant element of the MAS
(3) Soldier ant: they are in charge of defending the colony from
4. The scenario: A colony of Nothomyrmecia the attack of enemies.
macrops ants2 The eggs, which are put by the queen ant, are moved by the
Nothomyrmecia macrops is one of the most known and ancient
worker ants to a place with an appropriate temperature. The
ant species. It is believed that it is one of the first species and it
maggotss, which must be fed by the workers until they become
is characterized by having a very simple social organization.
adult, come from these eggs.
There are three types of adult ants, namely: the queen ant, the
These ants have a unique chemical sign used to identify them,
worker ant, and the soldier ant. Each ant type provides certain
and this characteristic permits other ants to identify whether they
funcionality to the society:
belong to the same specie and the type of ant it belongs to.
(1) Queen ant: the survival of the specie depends on the queen Furthermore, they have the capability of leaving other chemical
ant because it is responsible for the reproduction of the specie. signs in order to indicate the path to an area with food or to an
(2) Worker ant: they are in charge of collecting food, taking care enemy. They have antennas on the head that can be used for
of the eggs and keeping the maggotss in good temperature and communicating with other each other (i.e, communicating the
feeding condition. existence of enemies, asking for help, etc).
5. Results
In this section, the model obtained from the problem description
2
Characteristics from other species have also been included is introduced. First, the ontology used by each individual ant is
because they were considered to be interesting for studying the described. Then, the different agent's control modules are
social system exposed, and finally the social laws that coordinate the
multiagent system are presented. Figure 2 shows the hill; the survival of the specie; searching for an empty
categorisation for some of the most significant elements of the cell; collecting food; searching for food out of the ant
multiagent system according to Sowa's natural categories. hill; feeding maggotts; keeping maggotts safe; asking
for help; finding a new task to perform.
5.1 The multiagent system Defining the social objective: keeping the queen alive.
The model formalized in subsection 3.3.4 is particularized now
for the application domain chosen in this work. All the relevant 5.2 The ontology
aspects are intuitively described in the following lines: The main elements of the ontology modelled are the
Specifying the cell attributes that are part of the environment and the objects that can exist in it, the senses of an
environment: temperature and smelling. ant, its objectives, its strategies for achieving its goals, and the
Defining the different classes of objects that can be primitive actions that can make on the environment.
found in a cell: an ant, food, or a rock. 5.2.1 The environment and its objects
Specifying the state of the agent in terms of its The environment is formed by a set of cells in a rectangular
smelling, health, current strategies, and sensors value. structure. Each cell is one of the locations in two dimensions,
Defining the actions that can be performed on the and each cell has a determined temperature and has a specific
environment: move (direction); eat; grasp(object); smelling, which can be a mixture of different primitive
leave(object); put an egg; leave sign. smellings. A cell can contain at a particular moment an ant,
food, or an obstacle (e.g., a rock).
Defining the actions that can be performed on another
agent: touch (movement). An ant can be either adult or not, and it can be considered by
other ants as friend or enemy, attending to the specie it belongs
Defining the strategies: going to the ant hill; searching to. An adult ant, according to its social role can be queen,
for food inside the ant hill; searching for an empty worker, or soldier. The ant's condition is formed by its health, its
cell; going outside the ant hill; searching for food out safety sensation, and its stack of goals to achieve. The food is an
of the ant hill; carrying food to the ant hill; following object with a concrete energetic value. The rocks are a physical
a sign; searching for a maggott; warning another ant; obstacle that the ants must avoid. A part of the ontological
take a maggott to a good place. modelling of the objects that exists in this application domain
Defining the objectives of each agent: feeding itself; can be found in Figure 3. There, the ontology is edited by using
searching for food inside the ant hill; staying at the ant the ONTOIN tool mentioned in subsection 2.4.
Figure 3. Editing the ontology with the ONTOIN tool
5.2.2 The ant's senses is possible to indicate the path to an area with food or to an
enemy.
The ant has three sensors used for obtaining information from
the environment. These three sensors are: (2) Visual sensor: the sight is one of their less developed senses
so that its scope is shorter than that obtained with the smelling.
(1) Chemical sensor: it is used for identifying smellings in a
Despite this fact, it can be useful for detecting the other ants’
determined radius. The ants give off a unique smelling through
head movements. These head movements have a meaning and
which they can communicate to other ants the specie they
they are useful to tell another ant about the need of help for
belong to, the type of ant, and even the exact ant; it acts as an
attacking an enemy, collecting food or for other different tasks.
identification card. This can be useful for the same ant to know,
for instance, the path it has used by following the sign it (3) Temperature sensor: it is used as a reference when they look
previously left. Moreover, they are able to leave on purpose for places with appropriate temperatures to leave the food, the
other chemical signs attending to the task they perform so that it eggs or the maggotss.
5.2.3 The objectives, the strategies and the actions (food); (4) go_to (ant_hill); (5) search (safe_place); and (6) store
The objective of every ant is to maintain its level of energy and (food, safe_place) (it can imply to move to the 'safe_place').
security into an acceptable (pre-defined) range of values. The On the other hand, if the worker ant wants to follow another
strategy for keeping the nutritional level is based on detecting ant's chemical sign, the actions sequence would be the following
when its health condition has reached certain threshold and if so, one: (1) go_to (out_of_ant_hill); (2) smell (chemical_sign); (3)
proceed to search for new food. Concerning the secutrity issue, go_to (place_indicated_chemical_sign); (4) collect (food); (5)
the ant can take shelter in the ant hill or asking other ants for go_to (ant_hill); (6) search (safe_place); (7) store (food,
help; otherwise it will fight. safe_place).
The queen ant’s objective is the survival of the specie. Its All these processes belong to the ontology and must obey the
strategy consists in putting eggs periodically whenever she has a social laws defined for the MAS and reflected into the ontology
good health condition. as axioms. Each ant possesses an ontology which is all the static
The worker ants have two social objectives: (1) collecting food; knowledge the agent has about the MAS it belongs to. Each ant's
and (2) taking care of the maggots. In order to collect the food, ontology depends on its role, although ants with the same role
they can go out and searching for food or they can try to follow have the same ontology since the can perform the same
the sign of another worker ant which pursues the same goal. processes and they have the same goals. However, all the
Then, the food will be stored in the ant hill, particularly, in areas ontologies are based on a global MAS ontology which contains
with appropriate temperature for its conservation. For taking the overall functionality of the system and is (partially) reused
care of the maggotts, some worker ants will feed them. They for generating the particular ontologies for each type of ant.
can also be in charge of moving the eggs and the maggotts to
areas with the most appropriate temperature. Normally, an idle 6. Conclusions
worker ant will wander in the ant hill until finding a new task to In this work a multiagent system has been presented that
perform, or it will go out of the ant hill and search for food. attempts to partially simulate an ant colony. One of the
currently most significant tools for representing knowledge, the
The soldier ant’s goal is to keep the colony safe against ontologies, has been used for modelling the agents' internal
enemies.The strategy for achieving it consists in staying close to structure, so that our model can take advantage of the main
the ants it must defend. The degree of security an ant has benefits of the ontological representation, namely knowledge
depends on two factors: (a) the number of soldier ants that ant reusability and shareability. All the agents share the same
has around it; and (b) the number of enemies that can be vocabulary since their particular ontologies come from a
detected by both ants (i.e. the worker and the soldier one). common, global one so that they can directly share information
In the current design, the starting point is an ant colony which and distinctions. In this way, the groups of agents do not need to
has no intention of conquering other ant hills. However, the develop a common lexicon as it is needed in other approaches
possibility of the appearence of enemies from other ant species (see [22]).
has been reflected in the model. Then, an ant from the other The ontologies used for modelling the agents are based on the
species, which belongs to the environment, will have as its top-level categories established in [21] so that every single
objective to attack every ant from the rest of species. When entity that exists in the agent's world must belong to one of the
attacking, an ant can bite, and the soldier ants can also segregate twelve basic categories. Indeed, as it can be observed in Figure
a harmful acid. 2, the concepts that appear in our application domain correspond
The primitive actions an ant can perform over the environment to natural categories defined in the quoted work.
are: (1) moving in whatever direction; (2) eating food; (3) biting In particular, in the Artificial Society modelled here agents are
another ant; (4) grasping an object; (5) leaving an object; (6) assumed to possess a set of capabilities that guarantee their
segregating acid; (7) putting an egg; (8) moving the head; and ‘animated’ character. These are nutrition, perception,
(9) leaving a chemical sign. locomotion, desire and thought. In order to make these
capabilities operative by the agents, various types of control
5.2.4 Contribution for modelling the MAS
modules have been implemented. Such modules allow the
In this subsection, we explain the contribution of the ontology
agents to perform knowledge acquisition, to monitor their
for modelling the MAS through the exposition of an example.
situations for identifying objectives and to perform planning and
Let us suppose that we have a worker ant which has to collect
control as such. The complexity of such control modules will
food. Then, the ant has two possibilities: a) going out and
depend on the complexity of the application domain since the
searching for food by itself; b) going out and follow the
manner in which nutrition, perception, locomotion, desire and
chemical sign of another ant which pursues the same goal. Each
thought are implemented differs from an application domain to
agent is provided with an ontology which specifies all the
another. Furthermore, different domains might require different
actions it can perform and all its static knowledge about the
'psyches'.
world. For instance, this is a summary of the processes a worker
ant has in its ontology defined: go_to(place), search(object), An ontological framework that allows to build both
collect(object), smell(object), store(object, place). Therefore, inconsistency-free taxonomies and partonomies has been utilised
when a worker ant is at the ant hill and wants to search for food for modelling ants’ communications. The concepts for the
by itself, the sequence of actions can be roughly described as resulting ontologies are all grounded on the above mentioned
follows: (1) go_to (out_of_ant_hill); (2) search (food) (it can natural categories, and include for each ant its conceptualisation
imply to move to the place where the 'food' is); (3) collect about all the environment (and the objects it contains), its
senses, its objectives, its strategies and the actions to be taken.
To end, we have also described the social laws that must govern [15] Jennings, N.R., & Campos, J.R. (1997): Towards a social
the MAS designed in our work. The organizational rules used in level characterisation of socially responsible agents. IEEE
[25] can be considered similar to the social laws here used Proceedings on Software Engineering 144(1).
because they indicate the safety properties of the organization. [16] Minsky M. (1987): The Society of Mind, Picador, London.
[17] Newell, A. (1982): The knowledge level. Artificial
7. Acknowledgements Intelligence 18:82-127.
The second author is supported by the Fundación Séneca,
Centro para la Coordinación de la Investigación, through the [18] Paniagua-Arís, E., Martínez-Béjar, R., Palma-Méndez, J.T.,
FPI program. & Martín-Rubio, Fernando (2000): Towards a Teleo-Social
Agent Category. Third International Conference on
8. References Knowledge-Based Intelligent Information Engineering
Systems, Adelaide, Australia.
[1] Brooks R.A. (1991a).‘Intelligence without representation’,
Artificial Intelligence, No 47, p.p. 139-159. [19] Software Agents: An Overview. British
Telecommunications.
[2] Brooks R.A. (1991b). ‘Intelligence without reason’,
http://www.labs.bt.com/projects/agents/publish/papers/revi
Proceedings of the Twelfth International Joint Conference
ew1.htm
of Artificial Intelligence, Darling Harbour, Sydney,
Australia 24-30 August, Morgan Kaufmann, p.p. 569-595. [20] Sowa, J.F. (1995): Top-level ontological categories.
International Journal of Human-Computer Studies 43:669-
[3] Castelfranchi, C. (1998): Modelling social actions for AI
685.
agents. Artificial Intelligence 103: 157-182.
[21] Sowa, J.F. (2000): Knowledge Representation. Logical,
[4] Chandrasekaran, B., Josephson, J.R., & Benjamins, V.R.
Philosophical and Computational Foundations. Brooks-
(1999): What are ontlogies, and why do we need them.
Cole.
IEEE Intelligent Systems, 20-26.
[22] Steels, L (1997): Constructing and Sharing Perceptual
[5] Cyc Proyect, The. http://www.cyc.com/cyc-2-1/cover.html
Distinctions. Coupling of meaning creation and naming
[6] Duineveld, A.J., Stoter, R., Weiden. M.R., Kenepa, B., & games in software experiments. In: van Someren, M. and
Benjamins, V.R. (1999): WonderTools? A comparative G. Widmer (eds.) (1997) Proceedings of the European
study of ontological engineering tools. International Conference on Machine Learning. Springer-Verlag, Berlin.
Journal of Human-Computer Studies 52:1111-1133.
[23] Steels, L. (1995): When are robots intelligent autonomous
[7] Epstein J.M. & Axtell R. (1996): Growing Artificial agents? Journal of Robotics and Autonomous Systems. 15
Societies, MIT Press, Cam. (1995) 3-9
[8] Fernández-Breis, J.T. & Martínez-Béjar, R. (2000a): A [24] Studer, R., Benjamins, V.R., & Fensel, D. (1998):
web-based framework for integrating knowledge. Expert Knowledge engineering, principles and methods. Data and
Systems With Applications 18(4): 315-330. Knowledge Engineering 25:161-197.
[9] Fernández-Breis, J.T. & Martínez-Béjar, R. (2000b): A [25] Zambonelli, F., Jennings, N.R., Wooldridge, M. (2001).
cooperative tool for facilitating Knowledge Management. Organisational rules as an Abstraction for the Analysis and
Industrial Knowledge Management: A micro-level Design of Multi-Agent Systems. International Journal of
approach: 123-138. Software Engineering and Knowledge Engineering (to
[10] Fernández-Breis, J.T., Paniagua-Arís, E., & Martín-Rubio, appear)
F. (2000): Analysis of Multi Agent Systems for
representing human organisations (in Spanish). SEID'2000,
Orense, Spain.
[11] Fitoussi D. & Tennenholtz M. (2000): Choosing social
laws for multi-agent systems: Minimality and simplicity.
Artificial Intelligence 119: 61-101.
[12] Fridman Noy N. & Hafner C.D. 2000. Ontological
foundations for experimental science knowledge bases.
College of Computer Science. Northeastern University.
Boston, Massachusetts, USA.
[13] Goldspink C. (2000): Modelling social systems as complex:
Towards a social simulation meta-model.
http://www.soc.surrey.ac.uk/JASSS/3/2/1.html
[14] Gruber, T.R. (1993): A translation approach to portable
ontology specifications. Knowledge Acquisition Journal
5:199-220.