Motivation

Software Agents represent a really interesting paradigm. Based on what I read, I believe that this paradigm witnessed a surge in the early nineties, but was not adopted by the industry then. Right now we are witnessing a resurgence of this paradigm with the sincere hope this time its more widely accepted that it was the last time.

My exposure with Agents was a chance factor. I had to think of an idea for a project for the Mobile Computing Class. After a thorough literature search, I decided to implement an Agent based conference assistant, Rendezvous, though not fully realizing at that time that agents represent a much more stronger paradigm, rather than just as an implementation mechanism. I used JADE for our development and seeing the features of JADE, I was really impressed with what an Agent can give us. Thereafter I got a chance to implement a few other systems MagicWeaver, CoolAgentRS using JADE and this implemenation experience made me pose quite a few questions on the agents paradigm. This is where my research interest in agents started. UMBC runs an AgentWEB , through which I got to understand a lot of ideas prevailing in this domain.

In Fall 2001, I took a class, Engineering Multi-Agent Systems. This exposed to me a lot of questions about agents, what they are, what they can ideally do, coordination and communication among them, knowledge representation in these agents, Game theory applied to agents and agent applications and development environments. We are also incorporating Agents in ITTALKS , to explore the potential of agents to interpret smeantics of data expressed in DAML and perform interesting reasoning.

As I started work on Sensor Networks, I was initially geared towards core sensor network work but finally ended up using Agents for building the simulation platform. Infact the use of the notion of Agency did make my Masters Thesis all the more interesting.Agents do interest me a lot and I would welcome any oppurtunity to work on applied research aspects of these agents.

Area of Interest

I am interesting in the following aspects of agents:

1. Multi-Agent applications and architectures
2. Reasoning and Knowledge Representation among agents
3. Context-Aware Software Agents
4. FIPA specific issues like Communication, Coordination, Interoperability among software agents.
5. Use of the Agents Paradigm in building distributed components and how OPEN Agent Environment pose so many different challenges in realizing multi-agent systems.
6. How Agents or Autonomous software processes will be integrated in the Semantic Web. I do strongly believe that Semantic Web will be truly realizable and I would be interested in seeing what role software agents play in this realization. Infact, I do try to think a lot of times of how to embed agents in the next generation web.

Proposed/Accomplished Work

I have been involved in building a lot of agent-based systems and thats where I got all my ideas and notions of agency. The Multi-agent systems class gave me a clear picture into the research aspects of software agents and hence I became highly interested in the Communication and Coordination aspect of agents. I have had some indepth work experience in working with the internals of the agent platform like JADE and LEAP. Further, I have modified the Directory Facilitator (also an agent) in JADE to do XSB based reasoning for service lookup. The core systems where agents were used to drive the application where Rendezvous, CoolagentRS, ITTalks on Agentcities. MagicWeaver simulation platform was realized using the Agents paradigm but this was completely abstracted out for the sensor network designer. After attending the Agenticites workshop, I do tend to form the opinion that Agents in Open Environments would be really challenging to explore. In prototyping a particular service for Agentcities, there were so many issues that came up becuase of the openness of the environments.