LOGIC-BASED MOBILE AGENT FRAMEWORK

USING WEB TECHNOLOGIES

Shinichi Motomura, Takao Kawamura, Kazunori Sugahara

Tottori University

4–101, Koyama-Minami, Tottori 680–8552, JAPAN

Keywords:

Mobile agent, Logic Programming language, XML-RPC.

Abstract:

We have proposed Maglog which is a framework for mobile multi-agent systems. Maglog is based on Prolog,

and has the concept of ﬁeld. A ﬁeld is an object which can contain a knowledge base. With the concept

of ﬁeld, Maglog provides a simple and uniﬁed interface for 1)inter-agent communication, 2)agent migration

between computers, and 3)utilization of data and programs on computers. An agent migrates using HTTP as

the transport protocol and XML as the encoding format itself. In this paper, we present the implementation

of Maglog on Java environment, in detail. Since we have implemented both command-line shell and GUI for

Maglog, users can choose them for their needs. In addition, through XML-RPC interface for Maglog which we

have also implemented, other systems can easily utilize Maglog. As examples, we outline several applications

developed through XML-RPC interface.

1 INTRODUCTION

Multi-agent system is drawing attention as a structural

model for many software systems including distrib-

uted systems and artiﬁcial intelligence systems. In a

multi-agent system, a number of autonomous agents

cooperates mutually and achieves given tasks. Each

agent is generated according to a given task and can

have its own situation and operates under the sit-

uation. Situation consists of states and procedures

where both of them can be dynamically changed in

general. Therefore, it becomes necessary for the agent

to dynamically hold the states and procedures (here-

after we refer them as a knowledge base). More-

over, when a number of agents are cooperating, it is

necessary for them to share knowledge bases and to

conduct knowledge communications between agents.

In addition, mobility of agents is important in multi-

agent system because of not only reducing network

latency but simplifying architecture of software sys-

tems.

We have proposed Maglog (Kawamura et al., 2003;

Kawamura et al., 2004a) which is a framework for

mobile multi-agent systems. Maglog is based on Pro-

log, and has the concept of ﬁeld. A ﬁeld is an object

which can contain a knowledge base. With the con-

cept of ﬁeld, Maglog provides a simple and uniﬁed in-

terface for 1)inter-agent communication, 2)agent mi-

gration between computers, and 3)utilization of data

and programs on computers. In this paper, we present

the implementation of Maglog on Java environment,

in detail.

2 OVERVIEW OF MAGLOG

Figure 1: Overview of Maglog.

Figure 1 shows the overview of Maglog. An agent

runs as a thread in a process which we call an agent

server. Mobile agents of Maglog are written in Pro-

log. Agent servers have objects which hold Pro-

198

Motomura S., Kawamura T. and Sugahara K. (2005).

LOGIC-BASED MOBILE AGENT FRAMEWORK USING WEB TECHNOLOGIES.

In Proceedings of the First International Conference on Web Information Systems and Technologies, pages 198-201

DOI: 10.5220/0001228401980201

 SciTePress

log clauses. We call them ﬁelds. Builtin predicate

in(Goal, Field) is for evaluation of a goal in a

ﬁeld. An agent can enter a ﬁeld by this predicate. En-

tering a ﬁeld, an agent can utilize data and programs

in the ﬁeld. With the concept of ﬁeld, Maglog pro-

vides a simple and uniﬁed interface for 1)inter-agent

communication, 2)agent migration between comput-

ers, and 3)utilization of data and programs on com-

puters. An agent migrates using HTTP as the trans-

port protocol and XML as the encoding format itself.

2.1 Inter-agent Communication

Agents belonging to the same ﬁeld can be considered

of forming a group. The knowledge within the ﬁeld

is shared by the agents. Moreover, by changing the

knowledge within the ﬁeld, agents can inﬂuence the

actions of other agents.

An agent can communicate with other agents

synchronously or asynchronously by reading/writing

Prolog clauses from/into ﬁelds. Updating knowledge

base in a ﬁeld can be done by the following predi-

cates.

2.2 Migration

The second arguments of the following predicates can

be like FieldName@HostAddress.

in(Goal, Field)

fasserta(Clause, Field)

fassertz(Clause, Field)

fretract(Clause, Field)

fclause(Head, Body, Field)

If a host address speciﬁed, the agent will go to the

host and access the ﬁeld.

2.3 Dynamic Change of Behavior

An agent can change its behavior dynamically

through entering a ﬁeld. Figure 2 shows an exam-

ple. The execution of the goal print(’Hello!’)

sends the string “Hello!” to a printer when the agent

is in ﬁeldA, on the other hand, the same goal creates a

new window containing the string “Hello!” when the

agent is in ﬁeldB.

3 IMPLEMENTATION

We have implemented Maglog on Java environment

through extending PrologCaf

e which is a Prolog-to-

Java source-to-source translator system (Banbara and

Tamura, 1999). Both agents and after-mentioned sta-

tic ﬁelds are translated into *.java ﬁles with our Ma-

glog translator and then compiled into *.class ﬁles

with a Java compiler.

fieldA

fieldB

in(print(’Hello!’),

Field).

Field = fieldA

Field = fieldB

print(X) :-

...

print(X) :-

...

Hello!

window

agent

Hello!

Figure 2: Dynamic Change of agent’s behavior through en-

tering a ﬁeld.

Two functions for agent migration have been im-

plemented: a serialize function for agent and a RPC

function. These functions have been implemented by

XML-RPC to agent servers. We show the migratation

steps below.

1. An agent server encodes a agent to a XML docu-

ment.

2. The agent server gets a reference to a destination

agent server using XML-RPC.

3. The agent server invokes the destination agent

server’s RPC with the XML document. The RPC

uses HTTP as transfer protocol.

4. The invoked agent server decodes the agent from

a XML document and continues execution of the

agent.

5. The agent server makes up to a XML document

from the execution result and returns the XML doc-

ument.

3.1 Basic Components

In this section, we describe the Java implementation

of agent, agent server, and ﬁeld.

1. Agent

Figure 3: Structure of an agent.

Figure 3 shows the structure of an agent. An agent

executes predicates using Prolog Interpreter and

LOGIC-BASED MOBILE AGENT FRAMEWORK USING WEB TECHNOLOGIES

199

moves between agent servers. An agent has parent-

and-child structure and the child agent has the ad-

dress for returning to the parent agent as HomeAd-

dress.

2. Agent Server

Figure 4 shows the structure of an agent server. An

agent server creates an agent by createAgent

method. createAgent method creates an agent

from the agent repository, which contains classes

of predicates of agents, and pushes into the agent

scheduler. The agent scheduler has threads for

agents. When receiveAgent method is in-

voked, the MaglogAgentRemoteServer re-

ceives a agent via network, decodes it from XML,

and pushes into the agent scheduler. When

sendAgentmethod is invoked, the MaglogAgen-

tRemoteServer encodes a agent to XML, and sends

to a destination agent server via network. These

transfer protocols are HTTP.

Figure 4: Structure of an agent server.

3. Field

We have implemented two kinds of ﬁeld, dynamic

ﬁeld and static ﬁeld. Static ﬁelds are pre-compiled

predicates. Therefore agents cannot modify static

ﬁelds, however executions of predicates in static

ﬁelds are relatively fast. Executions of predicates in

dynamic ﬁelds are relatively slow however agents

can assert and retract clauses in dynamic ﬁelds.

3.2 Web Services Function

Agent Servers have Web Services are implemented

using XML-RPC. XML-RPC is a simple remote pro-

cedure calling protocol using XML as the encoding

format (Winer, 1998). For this function, it becomes

easy to use Maglog as a part of application.

Through XML-RPC, other systems can do the fol-

lowing operations.

• create and kill agents.

• create and delete dynamic ﬁelds.

• assert and retract clauses in dynamic ﬁelds.

• get a list of names of dynamic ﬁelds.

• get a list of IDs of agents.

Prolog clauses in return values and arguments of

requests are translated to data types of XML-RPC by

the agent server.

4 APPLICATIONS

We will outline several applications developed using

Maglog.

1. e-Learning System (Kawamura et al., 2004b)

An P2P-based e-Learning system has been built us-

ing Maglog. This e-Learning system has two dis-

tinguishing features. Firstly, it is based on P2P ar-

chitecture for scalability and robustness. Secondly,

each content in the system is not only data but an

agent so that it can mark user’s answers, tell the

correct answers, and show some extra information

without human instruction. Maglog plays an im-

portant role to realize the both features. Figure 5 is

a screen-shot of the user interface program. of this

e-Learning system, which is developed in Squeak

environment. This e-Learning system consists of

about 2,000 lines of Maglog code and about 4,000

lines of Squeak code.

Figure 5: The e-Learning system.

2. Schedule Arrangement System (Kinosita et al.,

2003)

This system, which has been developed using Ma-

glog, establishes and arranges meeting schedule

without human negotiations. Once a convener

convenes a meeting through the system, an agent

moves around the members of the meeting and ne-

gotiates with them automatically. This schedule

arrangement system consists of about 400 lines of

Maglog code and about 4,000 lines of Java code.

WEBIST 2005 - INTERNET COMPUTING

200

3. HECS System (Banbara et al., 2003)

Maglog is used for HECS (Heterogeneous Con-

straint Solving) system, which was supported in

part by IPA (The Information-technology Promo-

tion Agency) under grant of 2003 Exploratory Soft-

ware Project, to coordinate distributed solvers each

other.

5 RELATED WORKS

There are several mobile agent frameworks realized

as a set of class libraries for Java such as Aglets

(Lange and Oshima, 1998) and MobileSpaces (Satoh,

2000). The combination of one of them and a Prolog

interpreter/compiler written in Java such as NetPro-

log (de Carvalho et al., 1999) and Jinni (Tarau, 1999)

have some similarity to Maglog. The main difference

between the combination and Maglog is the class of

mobility. Their mobility is so-called weak mobility,

in which only its clause database is migrated. On

Maglog, all of the execution state including execu-

tion stack can be migrated (so-called strong mobility),

therefore agents on Maglog can backtrack and unify

variables during migration. That makes programs on

Maglog simple and understandable.

6 CONCLUSION

In this paper, we presented the implementation of

a mobile agent framework Maglog, in detail. With

the concept of ﬁeld, Maglog provides a simple and

uniﬁed interface for 1)inter-agent communication,

2)agent migration between computers, and 3)utiliza-

tion of data and programs on computers. And through

XML-RPC interface which we have implemented,

other systems can easily utilize Maglog. Through

XML-RPC interface, several applications have been

developed.

REFERENCES

Banbara, M. and Tamura, N. (1999). Translating a linear

logic programming language into Java. In M.Carro,

I.Dutra, et al., editors, Proceedings of the ICLP’99

Workshop on Parallelism and Implementation Tech-

nology for (Constraint) Logic Programming Lan-

guages, pages 19–39.

Banbara, M., Tamura, N., Inoue, K., Kawamura, T., and

Tamaki, H. (2003). Java implementation of a distrib-

uted constraint solving system. Exploratory software

project, Information-technology Promotion Agency

Japan.

de Carvalho, C. L., Pereira, E. C., and da Silva Julia, R. M.

(1999). Netprolog: A logic programming system for

the java virtual machine. In Proceedings of the 1st

International Conference on Enterprise Information

Systems, pages 591–598. Setubal, Portugal.

Kawamura, T., Kinoshita, S., and Sugahara, K. (2004a).

Implementation of a mobile agent framework on java

environment. In Gonzalez, T., editor, Proceedings

of the IASTED International Conference Parallel and

Distributed Computing and Systems, pages 589–593.

MIT, Cambridge, USA.

Kawamura, T., Kinoshita, S., and Sugahara, K. (2004b). A

mobile agent-based p2p e-learning system. In Gon-

zalez, T., editor, Proceedings of the IASTED Interna-

tional Conference Parallel and Distributed Comput-

ing and Systems, pages 873–877. MIT, Cambridge,

USA.

Kawamura, T., Kinoshita, S., Sugahara, K., and Kuwatani,

T. (2003). A logic-based framework for mobile multi-

agent systems. In Hexmoor, H., editor, Proceedings

of International Conference on Integration of Knowl-

edge Intensive Multi-Agent Systems, pages 754–759.

Boston, Massachusetts, USA.

Kinosita, S., Kawamura, T., and Sugahara, K. (2003). Mo-

bile agent based schedule arrangement system. In Pro-

ceedings of the 5th IEEE Hiroshima Student Sympo-

sium (HISS), pages 205–206.

Lange, D. B. and Oshima, M. (1998). Programming and

Deploying Java Mobile Agents with Aglets. Addison

Wesley.

Satoh, I. (2000). Mobilespaces: A framework for building

adaptive distributed applications using a hierarchical

mobile agent system. In Proceedings of IEEE Interna-

tional Conference on Distributed Computing Systems,

pages 161–168. IEEE Press.

Tarau, P. (1999). Inference and computation mobility with

jinni. In Apt, K., Marek, V., and Truszczynski, M.,

editors, The Logic Programming Paradigm: a 25 Year

Perspective, pages 33–48. Springer.

Winer, D. (1998). Xml-rpc speciﬁcation.

http://xmlrcp.com/spec.

LOGIC-BASED MOBILE AGENT FRAMEWORK USING WEB TECHNOLOGIES

201