IJCST Vol. 2, Issue 2, June 2011
ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)
Multi-Agent Based Disaster Management System: A Review
1
Swati Basak, 2Neelam Modanwal , 3Bireshwar Dass Mazumdar
1,2
Dept. of Computer Sc, BHU, Varanasi, India
3
SMS, Varanasi, India
Abstract
Disasters, like earthquakes, floods, etc. can cause large number
of deaths, injured and homeless. Appropriate responses are
needed in the form of allocating resources to handle the effects
of disasters. Multi-agent disaster system has been extensively
used in the different tasks of decentralized disaster problem
solving such as communication among agents, collective decision
making, cooperation, collaborative planning in large scale that
deals with uncertainty and conflicting information during disaster
management . In this paper, we analyze the current disaster
management and response systems, taking in consideration of
domain requirements, agent-based design methodology and
systems’ comparative view.
Keywords
Multi-agent system (MAS), Disaster Management (DM), Disaster
Response, coordination, collective decision making, cooperative
distributed problem solving, collaboration, coalition.
I. Introduction
Disasters, both natural and man-made, can strike anytime
or anywhere. There are two ways to overcome disasters: the
first is to prevent them from occurring, and second to have an
emergency system and plan of operation prior to the occurrence
of any crisis. In either approach, information technology plays
an important role in disaster management [16]. Disasters, like
earthquakes, floods, etc. can cause large number of deaths,
injured and homeless. Appropriate responses are needed in the
form of allocating resources to handle the effects of disasters.
Distributing resources with certain type of injuries to the nearest
and most adequate emergency stations and hospitals and allocating
transportation resources that are most suitable and that would
require the minimum time to get to the crisis location and from
the disaster area to designated treatment facilities [1].
An agent is a software program that acts flexibly on behalf of
its owner to achieve particular objectives. An agent must be an
autonomous, reactive and proactive with the quality of good
listener, analyzer and cooperative in nature as well as good
coordination, good communication, good collaboration, forming
good coalition and negotiation with other agents [27]. Agents in
a multi-agent system may have been designed and implemented
by different individuals, with different goals; therefore they may
not share common goals; hence, MAS are a system composed
of multiple interacting intelligent agents. It can be used to solve
problems which are difficult or impossible for an individual
agent or huge system to solve. Examples of problems which are
appropriate to multi-agent systems research include e-commerce,
disaster response, and modeling social structures, game theory,
energy management, autonomic computing, situation management
for BDI theory, electric power management, seismic emergency
management, power system, micro-grid system etc.
Multi-agent disaster system has been extensively used in the
different tasks of decentralized disaster problem solving such
as communication among agents, collective decision making,
cooperation, collaborative planning in large scale that deals
with uncertainty and conflicting information during disaster
management [27]. In detail, this type of disaster systems can
w w w. i j c s t. c o m
be viewed on information and knowledge fusion and take the
feedback from the existing agents for sensing, coordinating,
decision making and acting. It must be able to achieve these
objectives in environments in which: control is distributed;
uncertainty, ambiguity, imprecision; multiple agents with different
aims and objectives are present; and resources are limited and
vary during the system’s operation. Thus, it must exhibits the
properties of [17]: (i) more robust, interoperable, and priority
sensitive communications, (ii) better situational awareness and
common operating picture, (iii) improved decision support and
resource tracking, (iv) greater organizational agility, (v) better
engagement of the public.
In this paper, we analyze the current disaster management and
response systems, taking in consideration of domain requirements,
agent-based design methodology and systems’ comparative view.
A number of disaster response systems have been developed
based on multi-agents systems approach (such as: DrillSim [6],
DEFACTO [20], ALADDIN [29], RoboCup Rescue [30], and
FireGrid [9]) and more are being developed. A key aspect of such
multi-agent based response will be agent-assisted crisis actors (first
responder, managers, public) working together and also assist
the crisis actor in planning, and determining resources to use. At
last we conclude with the common limitations of current disaster
response systems and determine the comparative conclusion of
used disaster response systems.
II. Background of Disaster Management System
Disaster Management System can be viewed as four interrelated
sub-phases. The first is damage assessment, in which looses and
their magnitudes are identified. The second is needs assessment,
in which initially required response is identified. The third is
prioritization of response measures, in which required response
matches with available resources. If response demand is greater
than the current available resources, decision makers must
establish priorities or act for external resources. The fourth is
actual response, in which crisis resources are deployed, and
decisions are disseminated to responders and the population at
large. During the four sub-phases, crisis response activities face
challenge of reducing the influence of crises cause to society, the
economy, and the lives of individuals and communities and they
continuously adapt their behavior and make quick decisions to
tackle unpredicted events [17]. The domain of disaster system
is characterized as a virtual environment of required distributed
control, huge amount of data which are modular, decentralized,
changeable, ill-structured, and complex as well as uncertainty,
ambiguity with different objectives, and limited resources which
continually vary [2]. Design of disaster management system must be
include: (i) filtering and data fusion methods, (ii) decision-making
and machine learning methods for determining actions in response
to states, (iii) interaction mechanism to manage the interaction
between multiple actors and to model collective behavior, and (iv)
system architecture studies of different system organizations. The
problem of disaster can be solved by two issues:
A.Conventional Disaster Management System
It is also named as human-agent collective system where human
can make its own strategies for designing the model. This is the
International Journal of Computer Science and Technology 343
IJCST Vol. 2, Issue 2, June 2011
architecture for agent-based spatial decision support systems
(SDSS) for natural and man-made disasters; hence coordination,
communication and collaborative planning in large-scale in MAS
that deals with uncertainty and conflicting information during
crisis management, that means coordination and deployment of
multi-robots system or wireless sensors network are worked on
site. Human-Agent or Human-Robot interactions (HRI) during
search & rescue operations are the big security issues in agentbased systems for search and rescue missions, for agent-based
simulation systems, participatory simulations systems, for
behavioural modeling and simulation of rescue workers, and for
the development methods for agent based systems in emergency
management.
B. Agent Based Disaster Management System
It is also named as multi-agent system where agents simulate the
model for concluding the result. There are some following features
and characteristics of multi-agents for DM:
(i) Collective decision making: Collective decision making
exemplifies a ‘bounded rationality’ approach, has been advocated
by Herbert A. Simon; he said that collective decision making is
the aggregation of individuals’ information to generate global
solution [22]. This type of decision is used to serve a variety of
purposes from governance ruling to forecast planning. Internet
also hosts a suit of collective decision making system [12]; this
is the general example which use in daily life.
(ii) Coordination: Coordination is one of the key functionalities
needed to implement a multi-agent system. In a perfectly
coordinated system, agents will not work with other's agents to
achieve the sub-goals while attempting to achieve a common
goal; they will not need to explicitly communicate, as they will
be mutually predictable, perhaps by maintaining good internal
models of each other [24]. Coordination is used in the domain
of brokering and matchmaking in the E-Marketing system [3],
Disaster Management system [21], used in Mobile-Autonomous
system [22], etc.
(iii) Communication: Agents communicate in order to achieve
better the goals of themselves or of the society/system in which
they exist. The goals of communication might or might not be
known to the agents explicitly, depending on whether or not the
agents are goal-based. Communication can enable the agents to
coordinate their actions and behavior, resulting in systems that
are more coherent [12].
Speech-Act theory [28] describe the communication is by
interacting each communication primitive as an action that updates
the knowledge of an agent about all the aspects of the state.
AgentSpeak(L), Agent-0 and 3APL are the speech act based model
[4, 23, 33]. The most common used message passing languages
are KQML (Knowledge Query Markup Language) [34] and FIPAACL [11] which gives the protocol for information exchange,
independent of content syntax and facts of the application
domain.
(iv) Coalition: Coalition formations was addressed through
the cooperating groups of agents investigated within game
theory; however the game theoretic approach is centralized and
computationally intractable or an extensive form (i.e. sequential)
game, and then analyze them through Nash equilibrium to see how
the game would be played [10]. However, the Nash equilibrium
is often too weak because subgroups of agents can deviate in a
coordinated manner. The Strong Nash equilibrium is a solution
concept that guarantees more stability [7, 29].
(v) Cooperative Distributed Problem Solving (CDPS): CDPS
studies how a loosely coupled network of problem solvers can
344 International Journal of Computer Science and Technology
ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)
work together to solve problems that are beyond their individual
capabilities. Each problem-solving node in the network is capable
of sophisticated problem-solving and can work independently, but
the problems faced by the nodes cannot be completed without
cooperation. Cooperation is necessary because no single node
has sufficient expertise, resources, and information to solve a
problem [22].
It is used in the domain of Contract Net (CNET) is a high-level
protocol for achieving efficient cooperation through task sharing
in networks of communicating problem solvers [12, 22] as well as
used in Disaster System [1]. Another domain is distributed sensor
network, distributed constraint in Heuristic Search, task sharing
in heterogeneous system and Tower of Hanoi (TOH) problem
[12].
There are number of environments and domains existed where
MAS explore its features. Russell and Norvig suggest the following
classification of environment properties for multi-agent system
[8].
(i) Accessible versus inaccessible: An accessible environment
is one in which the agent can obtain complete, accurate, up-todate information about the environment's state. Most real-world
environments are not accessible in this sense.
(ii) Deterministic versus non-deterministic: A deterministic
environment is one in which any action has a single guaranteed
effect - there is no uncertainty about the state that will result
from performing an action. In contrast, a non-deterministic
environment contains uncertain, incomplete, imprecise and
ambiguous information [25].
(iii) Discrete versus continuous: An environment is discrete
if there are a fixed, finite number of actions and percepts in it
otherwise it is continuous.
(iv) Static versus dynamic: A static environment is one that can
be assumed to remain unchanged except by the performance of
actions by the agent. In contrast, a dynamic environment is one
that has other processes operating on it, and which hence changes
in ways beyond the agent's control. The physical world is a highly
dynamic environment, as is the Internet [5].
(v) Localized versus Geographically Distribution: An environment
is localized in which gathering of information and performs
the action in local area while in geographical environment,
information are collected from various sources of different
places and performing on them concurrently. This distribution is
also worked as the centralized and decentralized distribution of
information [31].
III.Types of Disaster Management and Response
System
MAS has been widely used in emergency situations resulting
from natural and human made disasters, such as flood, tsunami,
earthquake, terrorist attack, fire in building etc, represent complex
and dynamic environments with high level of uncertainty; hence
autonomous notification and situation reporting for disaster
management system will be done by multi-agent response system.
There are some following disaster response systems which are
generally used to design the disaster system:
A. DrillSim [6, 21]
DrillSim is a user-centric simulation environment for testing IT
solutions. The purpose of DrillSim is to play out a disaster response
activity where agents might be either computer agents or real
people playing diverse roles (first responders, crisis managers,
experts, etc.).
w w w. i j c s t. c o m
ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)
IJCST Vol. 2, Issue 2, June 2011
1. Architecture
DrillSim is a multi-agent simulation and modeling system.
DrillSim is based on scalable architecture (O (100,000) agents),
and is extended by plug-and-play capability. System components
include I/O interfaces, simulation engine, data management
module, database server, and the Virtual Reality/Augmented
Reality modules.
1. Architecture
DEFACTO is a multi-agent simulation and modeling system based
on Machinetta proxy architecture. Architecture of DEFACTO is
scalable (O (10,000) agents) and flexible. DEFACTO consists of
simulator, 3D omni-viewer, Machinetta proxy based teamwork
infrastructure, and analysis tool to analyze the impact of teamwork
interaction strategies.
2. Methodology
Each agent has a role and a profile (age, cognitive abilities, health,
and knowledge). Simulation scenarios are created by binding roles
and profiles to agents. DrillSim has modeled agent behavior as a
discrete process where agents alternate between sleep and awake
states. Agents wake up and take some action every t time units. For
this purpose, an agent acquires awareness of the world around it,
transforms the acquired data into information, and makes decisions
based on this information using recurrent neural network.
2. Methodology
DEFACTO has modeled agent in proxy team formation
(Machinetta). Machinetta proxies are responsible for transferof-control over a decision, managing local team beliefs,
communication between proxies, communication between proxy
and a team member, coordination, and task allocation for the team.
Each proxy provides all transfer-of-control strategy options. One
of strategy options is selected based on current situation and agent
role. An optimal transfer-of-control strategy balances the risk of
high quality decision made by human against the risk of costs
incurred due to a delay in getting the decision from agent; hence,
each team implements team-oriented plans which describe joint
activities to be performed.
3. Features
• System allows testing of IT solutions in the context of the
simulated response activity to study the effectiveness of the
solutions.
• System helps understanding the response activity.
• System integrates with other simulators (e.g., communication
simulators, crisis simulators).
• Clear interfaces between information processing stages.
• Scenario based agent interaction (Q Language).
• Agents involved in information flow.
• Human can control and communicate with agents.
• Calibration of agent response models and metrics via running
activity in simulated and real worlds.
• System keeps global log of every event, information exchange,
and decision. In addition agents keep an individual consistent
local log.
• Crisis real-time response support and training system.
4. Limitations
• System architecture does not support fault-tolerance.
• System does not support high level control. Future actions are
based on local agent behavior (operational level), permitting
agents to execute undesirable action which leads to miss
common operations goal.
• System does not support adaptive planning.
• The simulation engine includes the simulated geographic
space, the evacuation scenario, and the agents. By such design
approach, the switching to another implementation requires
considerable reworks.
• Offline agent learning; agents need to learn about new roles,
and information variables before scenario execution.
• Agent presents limited configurability in terms of decision,
motion and health models, because their characteristics can
be specified only through the hard-coded models.
B. Demonstrating Effective Flexible Agent Coordination
of Teams through Omnipresence (DEFACTO) [20, 32]
DEFACTO is a user centric system which incorporates 3D
visualization omni-viewer, and human-interaction reasoning
into a unique high fidelity system. Human-interaction allows
responders to interact with the coordinating agent team in a
complex environment, in which the responder can gain experience
that will be applicable in the real world.
w w w. i j c s t. c o m
3. Features
• Improved situational awareness via interactive omniviewer.
• Improved team performance through flexible human-agent
interaction strategies.
• System allows transferring control from human to agents’
team using team-level strategies.
• Conflict resolution algorithms.
• System divides global goal (strategic level) into sub-goals
(tactical level) represented as joint intentions; which are
executed via agent team members (operational level).
• Coordination, collaboration, and task allocation between
agent team members.
• Calibration of human-agent transfer-of-control strategies
• Crisis real-time response support and training system
4. Limitations
• System does not support fault-tolerance.
• Building 3D model for omni-viewer can require months or
even years of manual modeling efforts.
• Agents need high bandwidth communication channels to
communicate.
• Agent presents limited configurability in terms of agent
profile and scenarios.
• System does not support adaptive planning.
• System does not provide learning from experience
strategies.
C. Autonomous Learning Agents for Decentralized Data
and Information Systems (ALADDIN) [2, 13]
ALADDIN is a user centric system, which aims to model,
design, and build decentralized systems that can bring together
information from variety of heterogeneous sources in order to
take informed action. For that goal, ALADDIN is considering
different aspects such as data fusion, decision making, machine
learning, and system architecture.
1. Architecture
ALADDIN architecture is based on High Level Architecture
(HLA) standard. ALADDIN organizes the simulator software in
four different layers: (i) Simulation Model Layer, (ii) Simulation
International Journal of Computer Science and Technology 345
IJCST Vol. 2, Issue 2, June 2011
ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)
Components Layer, (iii) Discrete Event Simulation Layer, and
(vi) Distributed Discrete Event Simulation Layer. Simulation
Model Layer is the layer where the simulation model is defined
through the declaration of the agents involved in the simulation.
ALADDIN is scalable and reusable system working through
decentralized simulation framework.
to the situation to be executed. Agents coordinate with each others
to explore crisis space to find civilians. Agents’ motion paths
are evaluated through methods for hierarchical real-time path
planning. Agents predict the life-time of found civilians, and
collaborate to optimize rescue actions sequence using genetic
algorithms.
2. Methodology
System is composed of autonomous, reactive, and proactive agents.
Agents can sense, act and interact in order to achieve individual and
collective goals. Agents are grouped into coalitions and assigned
to specific task (operational level). Agents collaborate with each
others based on multi-dimensional trust and reputation model to
achieve global goal. Tasks are assigned to agents’ teams using
optimization technique based on neural network. Then, agents’
teams bid for resources to manage with unexpected resource
allocation situations [26].
3. Features
• Decentralized control.
• Prediction of hazard material spread (fire spread).
• System adopts high level plan (strategic level plan).
• System adopts space-exploration techniques.
• Improved situational awareness using sensors network.
• Prediction for the civilian’s life time via machine learning.
• Minimizing sequence fluctuations via genetic algorithms.
• Dynamic agent role allocation.
• Calibration of agent teams behavior.
• Crisis real-time response support system.
3. Features [14]
• System is based on decentralized architecture and real-time
response support system.
• Machine learning algorithms and control applied at the
strategic, operational and tactical levels with minimal agents
communication.
• Improved situational awareness via sensors network.
• System divides global goal into sub-goals represented as subgraphs, which are executed through agent actions (operational
level).
• System adapts to environmental changes via sensors
network.
• System involves adaptive on-line decision making
algorithms.
• System involves auctions to make agents’ coalition.
• System involves data fusion techniques.
• System adopts inference and prediction to predict agent future
events.
• Flexibility and reusability of agents.
4. Limitations
•
System architecture does not support fault-tolerance.
•
System does not support plug-and-play capability.
•
System lacks calibration tools of agent behavior.
•
System lacks powerful user interface.
D. RoboCup Rescue (ResQ Freiburg Project) [18,30]
RoboCup Rescue is a user-centric large-scale simulation for urbansearch and rescue. The main design goal of RoboCup Rescue is
to enable rescue teams to effectively cooperate despite sensing
and communication limitations.
1. Architecture
RoboCup Rescue is a multi-agent simulation and modeling System.
System components include simulation engine, knowledge base
of agent relations, debugging tools, and data mining software for
task evaluation. The RoboCup Simulation league is divided into
two subunits, (i) Agent Simulation, and (ii) Virtual Robots. Agent
simulation platform currently runs a kernel which connects Traffic
simulator, Fire simulator, and Civilian simulator. While, Virtual
robot is based on Urban Search and Rescue Simulation.
2. Methodology
The basic task of agents in RoboCup Rescue is to collect, store,
and evaluate information. Then agents choose best actions fitting
346 International Journal of Computer Science and Technology
4. Limitations
• System architecture does not support fault-tolerance.
• System does not support plug-and-play capability.
• Plans are not adapted to situation changes.
• Agents lack reusability.
E. FireGrid [9, 15]
FireGrid is a task-centric collaborative community to pursue
research for developing realtime response systems using the Grid.
FireGrid addresses response process in the built environment,
where sensor grids in large-scale buildings are linked to super-real
time grid-based simulations.
1. Architecture
FireGrid is based on task-centric I-X agent architecture. FireGrid
integrates several core technologies such as: (i) fire and structural
models, (ii) wireless sensors in extreme conditions with adaptive
routing algorithms, (iii) grid computing which involves sensorguided computations, and mining of data streams for key events,
and finally (v) command-and-control using knowledge-based
Hierarchical Task Network planning techniques with user
guidance.
2. Methodology
All system components are integrated in the command-andcontrol (C2) task. The C2 task can be defined as the exercise
of authority and direction over available resources towards the
accomplishment of some objectives. The C2 process consists of
repeated cycles of a number of subtasks similar to tasks adopted
in DrillSim agent behavior model.
3. Features
• Grid architecture for distributed computation.
• Improved situational awareness using sensors network.
• Self-Configuring sensors network.
• System supports plug-and-play capability.
• System enables high level plan.
• System supports agent safety and security.
• Crisis real-time response support system.
4. Limitations
• System architecture does not support fault-tolerance.
• Generated plans are not adapted to situation changes.
• Agent presents limited configurability in terms of decision,
w w w. i j c s t. c o m
ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)
•
•
motion and health models.
Calibration of system is valid for simple scenarios only.
System lacks flexibility and reusability of agents.
IV. Analysis and Comparative View of mentioned
systems
DrillSim, DEFACTO, FireGrid, and ALADDIN systems are
limited to study emigration crises. While, RoboCup Rescue is
focusing on urban search and rescue operations. DrillSim and
ALADDIN systems are for matching domain requirements; in which
DrillSim is designed to study information technologies metrics in
disaster response, and ALADDIN is designed to study different
agent architectures in response operations. In addition, current
systems had focused on roughly supporting response activities
with small interest on improving the effectiveness of response
operations. System development should take in consideration
domain requirements to increase response effectiveness.
V. Conclusion
The paper reviews the various aspects of agent oriented framework
of software product. DrillSim, DEFACTO, FireGrid, and
ALADDIN systems are limited to study emigration crises while
RoboCup Rescue is focusing on urban search and rescue operations.
Although these details are available in the concerned literature but
joint view makes more appealing and comprehensive.
References
[1] ABEER EL-KORANY, KHALED EL-BAHNASY, “A multiagent Cooperative Model for Crisis Management System”,
7th WSEAS Int. Conf. on ARTIFICIAL INTELLIGENCE,
KNOWLEDGE ENGINEERING and DATA BASES
(AIKED'08), University of Cambridge, UK, Feb 20-22,
2008.
[2] Adams, M. Field, E. Gelenbe, D. J. Hand, “The Aladdin
Project: Intelligent Agents for Disaster Management - IARP/
EURON Workshop on Robotics for Risky Interventions and
Environmental Surveillance”, 2008.
[3] A. Jadbabaie, J. Lin, A. S. Morse, “Coordination of Groups of
Mobile Autonomous Agents Using Nearest Neighbor Rules”,
2002.
[4] A. S. Rao, “AgentSpeak(L): BDI agents speak out in a logical
computable language” Agents Breaking Away:Proceedings of
the Seventh European Workshop on Modelling Autonomous
Agents in a Multi-Agent World(MAAMAW-96) 42–
55,1996.
[5] Avgoustinos Filippoupolitis, Georgios Loukas, Stelios
Timotheou, Nikolaos Dimakis, Erol Gelenbe, “
Emergency Response Systems for Disaster Management in
Buildings”.
[6] B a l a s u b r a m a n i a n ,
M a s s a g u e r,
Mehrotra,
Venkatasubramanian, “DrillSim: A Simulation Framework
for Emergency Response Drills”; Proc. of ISCRAM, 2006.
[7] B Douglas Bernheim, Bezalel Peleg, Michael D Whinston,
“Coalition-proof Nash equilibria: I concepts; Journal of
Economic Theory, 42(1):1-12, June 1987.
[8] B Douglas Bernheim, Michael D Whinston, “Coalitionproof Nash equilibria: II applications", Journal of Economic
Theory, 42(1):13-29, June 1987.
[9] Berry, D., Usmani, “A. FireGrid: Integrated Emergency
Response and Fire Safety Engineering for the Future Built
Environment”; UK e-Science Programme All Hands Meeting,
Nottingham, UK, Sept. 19-22, 2005.
[10]D Kuwabara, T Ishida, N Osato, “AgentTalk: coordination
w w w. i j c s t. c o m
IJCST Vol. 2, Issue 2, June 2011
protocol description for multiagent systems” Proceedings of
the First International Conference on Multi-Agent Systems
(ICMAS-95) 455 1995.
[11] FIPA, ‘‘Specification Part 2:Agent Communication
Language,’’ The text refers to the specification dated 23
October 1997.
[12]Gerhard Weiss, “Multi-Agent Systems – A Modern Approach
to Distributed Artificial Intelligence”, Reprint: 2001,
Massachusetts Institute of Technology; pp. 83, 87, 88, 125,
127, 129, 242, 356, 360.
[13]Gianni, Georgios Loukas, Erol Gelenbe, “A Simulation
Framework for the Investigation of Adaptive Behaviors in
Largely Populated Building Evacuation Scenarios”; The
International Workshop on "Organized Adaptation in MultiAgent Systems, at AAMAS 2008.
[14][Online] Available: http://www.ecs.soton.ac.uk/research/
projects/aladdin.
[15]I-Rescue project, [Online] Available: http://www.i-rescue.
org/
[16]Jintae Lee, Tung Bui, “A Template-based Methodology for
Disaster Management Information Systems”, proceedings
of the 33rd Hawaii international Conference on System
Sciences, 2000.
[17]Khaled M. Khalil, M. Abdel-Aziz, Taymour T. Nazmy,
Abdel-Badeeh M. Salem, “Multi-Agent Crisis Response
systems – Design Requirements and Analysis of Current
Systems”.
[18]Kleiner, B. Steder, C. Dornhege, D. Höfer, “RoboCupRescue
- Robot League Team RescueRobots Freiburg (Germany)”;
RoboCup (Osaka, Japan), 2005.
[19]Ku Ruhana Ku-Mahamud, Norita Md Norwawi, Norliza
Katuk, Safaai Deris, “Autonomous Notification and Situation
Reporting for Flood Disaster Management”; The research
is financed by MOSTI, IRPA grant No. 04-02-07-10040EAR.
[20]Marecki, Nathan Schurr, Milind Tambe, “Agent-based
Simulations for Disaster Rescue Using the DEFACTO
Coordination System”; Emergent Information Technologies
and Enabling Policies for Counter Terrorism, 2005.
[21]Massaguer, Vidhya Balasubramanian, Sharad Mehrotra,
Nalini Venkatasubramanian, “MultiAgent Simulation of
Disaster Response”; AAMAS’06 May 8–12 2006, Japan.
[22]Matthias Klusch, Sycara, “Brokering and Matchmaking for
Coordination of Agent Societies: A Survey”.
[23]M Dastani, Birna van Riemsdijk, Frank Dignum, J J Meyer
,“A Pogramming Language for Cognitive Agents Goal
Directed 3APL”, ACM 200X.
[24]Michael Wooldridge, “An introduction to Multi-Agent
Systems”, Reprint: 2002, JOHN WILEY & SONS, LTD;
pp. 164, 168, 170, 175, 194, 248, 319.
[25]Muhammad Aslam, Muhammad Tariq Pervez, Syed
Shah Muhammad, Seemal Mushtaq, Ana Maria Martinez
Enriquez, “FMSIND: A Framework of Multi-Agent Systems
Interaction during Natural Disaster”; Journal of American
Science, 2010.
[26]Nick Jennings, Gopal Ramchurn, Mair Allen-Williams, Raj
Dash, Partha Dutta, Alex Rogers, Ioannis Vetsikas; “The
ALADDIN Project: Agent Technology to the Rescue”.
[27]N.M. Adams M. Field, E. Gelenbe, D.J. Hand, N.R. Jennings,
D.S. Leslie, D. Nicholson, S.D. Ramchurn, S.J. Roberts,
A. Rogers, “The ALADDIN Project: Intelligent Agents for
Disaster Management ”.
[28]P.R. Cohen, C.R. Perrault, ‘‘Elements of a plan based theory
International Journal of Computer Science and Technology 347
IJCST Vol. 2, Issue 2, June 2011
ISSN : 2229-4333(Print) | ISSN : 0976-8491(Online)
of speech acts,’’ Cognitive Science, vol. 3, pp. 177-212,
1979.
[29]R Aumann. Acceptable points in general cooperative n-person
games. volume IVof Contributions to the Theory of Games,
Princeton University Press, 1959.
[30]RoboCup project, [Online] Available: http://www.robocup.
org/
[31]Scalem M, Bandyopadhyay S, Sircar A K, “An Approach
towards a Decentrralised Disaster Management Information
Network ”.
[32]Schurr, Janusz Marecki, Tambe, J. P. Lewis, “The Future
of Disaster Response: Humans Working with Multiagent
Teams using DEFACTO”; AAAI Spring Symposium on "AI
Technologies for Homeland Security", 2005.
[33]S. Russell and P. Norvig, “Artificial Intelligence: A Modern
Approach”, Prentice-Hall, 1995.
[34]Y Labrou , J Mayfield and T Finin “KQML as an agent
communication language” “Software Agents” ,MIT
Press,Cambridge ,1995.
348 International Journal of Computer Science and Technology
w w w. i j c s t. c o m