Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h - 2012 1

\ISSN 2229-5518

A Survey on Applications of Mobile Agents in


P Karthikeyan, Dr. E. Sathiyamoorthy

ABSTRACT- A mobile agent is a sof tw are program that can transport its state f rom one environment to another, w ith its data intact, and be capable of perf orming operations appropriately in the new environment. The use of mobile agent is a rising research f ield w hich has got a broad application in f uture. In this paper w e examine the characteristics and types of the mobile agents. We also examine applications of mobile agents in the f ield of E-Business.

Inde x Terms— Mobile Agents- E-Co mmerce Mult i-Agent

—————————— ——————————


n agent is an autonomous entity which observes and acts upon an environment and directs its activity towards achieving goals . Their behaviors were res tricted and have predetermined activ i-
ties . Now with a ris ing growth in technology agents have b ecome
mobile agents .
Mobile agent technologies had changed the way we live and work. It offers a new computing paradigm in which a program, in the form of a software agent, can s uspend its execution on a host computer, trans fer its elf to another agent-enabled hos t on the network, and resume execution on the new hos t. Nwana [1] class ify agents into the following types .
- Collaborative
- Interface,
- Mobile
- Information
- Reactive
- Hybrid
- Intelligent.
Some ess ential characteris tic features of the mobile agents are dis- cussed by Jeffrey M.Bradshaw [2],

The bas ic property of a mobile ag ent is to act on behalf of its us ers by moving across the network from host to host and bring back the result to the us er.

Agents bring the program clos er to the resources .

Agents have the capability of learning, cooperation with other agents and mobility.

Mobile agents migrates the computation to the data ins tead

———— ——— ——— ——— ———

P Karthikeyan is Assistant Professor and pursuing Phd in School of I n- formation Technology & Engineering in VIT University, India,


Dr E Sathiyamoorthy is Associate Professor in School of Information Tech-

nology & Engineering in VIT University, India,

E-mail: .

of the data to the computation.

It us es limited memory s pace and CPU cons umption.

Agents offers less usage of bandwidth by migrating itself to s erver and returns only after the process is over and hence also reducing network load and traffic.

Mobile agents are dynamic in nature as they can work on

various heterogeneous execution env ironments .

They are both autonomous as well as asynchronous .

The mobile agents have the ability to react dynamic ally in unfavourable circums tances s o that it can overcome faulty behaviour during their execution in complex distributed sys tems and hence poss ess the feature of robustness .

They provide convenient and eas ier development paradigm.


Software agents can be class ified into two types bas ed on their mobility. Stationary agents are those that do not move over the networks and executes only on the system on which it begins execution whereas , mobile agents move across the network to carry on its tas k. Stephen Hagg [5] suggests four vital types of intelligent mobile agents bas ed on the tas k they perform,
1. Shopping bots
Shopping bots are agents which ass ist us ers for shopping products and getting services online. They make the shopping process easy for the buyers by getting info r- mation on various product items and services and acknowl- edge the final res ult back to the one who dispatched it. They are called as s hopping bots because the word “bot” refers to robot. So these agents ’ acts like buyer’s pers onal robots that do his work. These can be mainly used for most for buying the products on trade and market alon g with electronic and other one s ize fits all products .
2. Us er or personal agents .
Us er agents are intelligent agents that perform the

IJSER © 2012

http ://

Inte rnatio nal Jo urnal o f Sc ie ntific & Eng inee ring Re se arc h Vo lume 3, Issue 3 , Marc h - 2012 2

\ISSN 2229-5518

tas ks for us er on their behalf. Some of the tas ks performed by user agents are checking of e-mail, play as your oppo- nent in games , gathering of information, automatic filling of forms , scanning of web pages , s earching of jobs etc.
3. Monitoring-and-s urveillance or predictive agents .
Predictive agents are intelligent agents that fore- s ee the changes of equipment (mainly computer s ystem) by keeping watch over them. If there occurs any changes (changes in web page content) , anomalies , defects , ma l- functioning or any other absurd behavior over thos e equip- ments , the agents reports back for action to be taken. The agents may also monitor company inventory levels , keep watch on competitors ' prices by monitoring the web page (eg. Web watch) and pass them on back to the company, etc.
4. Data mining agents .
Data mining agents work on discovering informa- tion from varied s ources . The us er can sort through this in- formation in order to find whatever information they are s eeking. Once the information is got from different sources we can take actions on it. Patterns of information can be ca- tegorized into different class es by thes e agents . They can also detect major shifts in trends and can find out the pres- ence of new data and alert you. Thes e data mining agents can discover a downfall in manufacturing industry of a sp e- cific product in economy. After dis covering, the indus tries can make intelligent decis ions based on the trans mitted in- formation.


Multi Agent in e-commerce provides a new way to conduct various types of bus iness s uch as B2B, B2C and C2C. [4]. Multi Agent systems in E-commerce can be clas- s ified into three categories namely, shopping agents , s ales- man agents and auction agents .


These Agents make purchas es in e-marketplaces on behalf of their owner according to user defined needs . A typical shopping agent may compare features of different products by vis iting several online s tores and report the best choice to its owner. Since the agent’s moves to the source of information, the overhead of repeatedly trans fe r- ring potentially large amounts of information over a ne t- work is eliminated. Example of s hopping agents is ShopBot [7] which works in two phases , wherein in the firs t phas e it creates a learning agent which gathers a vendor description for each merchant and in consequent comparison shopping phas e it uses the information collected in the learning phase and decides which store offer the best price for a given product.


These sales man Agents behave like a traveling s a-
les man who vis its customers to s ell his products . This mo d- el of e-commerce us es a supplier driven marketplace and is particularly attractive for products with short life time. A supplier creates and dispatches an Multi-Agent to poten- tial buyers by giving it a lis t of s ites to vis it. A sys tem imple menting s ales man agents is Firefly where it uses col- laborative filtering mechanis m for recommendation the given product [6].


MIT Media lab’s Kas bah [6] is an example of customer to customer multi-agent s ystem. Us ing this agent us ers who wishes to buy or s ell an item will create an agent by speci- fying the guidelines and moves it to the centralized market place. Thes e multi-agent proactively seek out potential buyer and seller and negotiate with them on behalf of the us er bas ed on the user-specific cons traints such as initial price, final price, data of completion and Incremental value.


Each of the above mobile agent types s erves varying purpose based on users needs . If a buyer wants to shop online and s earch for s imilar products in more than a s ite by negotiat- ing then it can use any one of the E-commerce agent types bas ed on the needs . It reduces user’s time and gives regular alerts for the buyer. The pers onal agents are said to take care of us er’s tas ks and it greatly provides p ersonal ass is tance and of great help to all Internet surfers . Remote monitoring and filtering are us ed to monitor computer sys tems and forecas t any changes prior to it become worse. A data mining agent is us ed to find in- formation from lots of different sou rces for user. All thes e types of agents greatly reduce the effort of humans and are becoming popular day to day. They are widely used and deployed in rich and unlimited domain fields .


[1] H. S. Nwana, Software Agents : An Overview, Knowledge
Engineering Review, Vol.11, No 3, pp.1-40, 1996.
[2] Jefrey M.Bradshaw, An Introduction to Software Agents , Software Agents , MIT Press Book, 1997.
[3] Danny B. Lange and Mits uru Oshima, "Seven Good Re a
s ons for Mobile Agents ", Communications of ACM , vol.
42, no. 3, March 1999.
[4] Abdelkader outtagarts , "Mobile Agent-based applications : a Survey ", International journal of computer s cience and network s ecurity, vol.9, No.11, pp- 331- 339 , 2009
[5] Stephen Haag. "Management Informatio n Sys tems for the in-
formation Age”, Pages 224-228, 2006.
[6] Pattie Maes , Robert H. Guttman and Alexandros G. Mou kas , "Agents That Buy and Sell ", Communications of ACM , vol. 42, no. 3, pp. 81 - 91, March 1999.
[7] Robert B.Doorenbos A Scalable comparis on -Shopping
agent for world wide web, page 39-48, 2007.

IJSER © 2012

http ://