Dimple Juneja

A Thorough Insight into Theoretical and Practical Developments in MultiAgent Systems

Multiagent systems have been a fascination for research community and are often seen as an intelligent solution to many complex real world problems. Researchers have been active in the domain since last three decades and many developments pertaining to theoretical design and practical developments of multiagent systems are worth appreciating. The growth in MAS is multidirectional ranging from conceptual ideas to practical implementations and from the wide range of applications; it appears that multiagent systems are proving to be universal. The paper presents a concise survey of developments in MAS highlighting the important contributions in the field and also questions the universal applicability of agents.

Design of an Intelligent and Adaptive Mapping Mechanism for Multiagent Interface

The main intent of this work is to propose an intelligent interface that facilitates agent interaction in homogenous as well as heterogeneous ontologies. Literature survey indicates that there exist mapping mechanisms serving well for homogenous domains but very few researchers have made an attempt to propose a mapping interface for heterogeneous domains that is possessed with learning abilities and is therefore adaptive by nature. This work uniquely contributes towards the future vision of an intelligent and adaptive mapping mechanism that not only overcomes the curse of already existing mapping mechanisms but also is time efficient. The performance of the proposed strategy has been evaluated and compared with the existing strategies in the related fields and the results are found be to be competitive.

A multi-agent hybrid protocol for data fusion and data aggregation in non-deterministic wireless sensor networks

Data fusion deals with collaborative in-network processing and gathers relatively accurate information about the events in the environment. Conventional data fusion algorithms when assisted with mobile agent technology shifts computationally intensive tasks to these intelligent units thereby increasing the lifetime of the network. There exist mobile agent based event driven protocols for accumulating and forwarding the information to the sink(base station) eg: Tree-based Itinerary Design [4]. Usually, such protocols either deploy value-based fusion or decision-based fusion but very few are using both at the same time. Moreover, use of multi-agent systems(MAS) in such protocols is still in its infancy. The focus of this work is thus to propose a multi-agent hybrid protocol exploiting the benefits of both value and decision fusion by performing aggregation at the source level in a clustered WSN.

An Agent-Based Solution to Energy Sink-Hole Problem in Flat Wireless Sensor Networks

Repeated and continuous transmission of data to the sink leads to energy loss in all the nodes in case of flat WSN. Especially, depletion of energy is highly acute in case of nodes that are near to the sink. Conventionally known as energy sink-hole problem, it causes early failure of the network even when there is a substantial amount of residual energy left in it. Though the research fraternity has been continuously addressing this problem and even has provided various solutions to deal with it, the use of mobile agents to meet the above-stated problem is still in its infancy. The paper proposes a mobile agent-based solution for solving energy sink-hole problem. The proposed solution aims to extend the network life by reducing redundant data being passed to the nodes near to the sink thereby reducing the load and saving battery life. The algorithm is implemented using aglets and the analytical results show significant improvement in the network lifetime.

Adaptive and automated fault-tolerance for multi-agent systems

Fault tolerance in MASs is usually implemented with dynamic replication strategy, which demands real time replication of all agents. This adds to unnecessary replicas, more cost and also increases system complexity. Therefore the challenge of balanced replication is still prevailing. The major thrust of this work is to overcome the mentioned challenge and hence provide an equilibrium between active and passive replication of agents. The proposed framework relies on the fact that only critical agents need to be replicated, wherein criticality of an agent is assessed on the basis of its cardinality. The proposed framework has been implemented in JADE and provides promising results. The solution proposed adapts well to the transient environment situations of the agents and provides fault tolerance by making optimum utilization of the system resources.

A Clustering-Based Generic Interaction Protocol for Multiagent Systems

The paper proposes a clustering based Generic Interaction Protocol for Multiagent Systems (GIPMAS) that exploits clustering methodology for establishing interaction among agents operating in a network of multiagent systems. GIPMAS is a hierarchical protocol that supports intelligent formation of clusters and dynamic election of cluster head and executive cluster head as well. It also describes a recovery mechanism in case cluster head relocates from its respective cluster.

Hybrid data fusion using software agents for an event driven application of wireless sensor networks

Data Fusion using software agents in WSNs is a naive area of interest for researchers belonging to the community of sensors as well as software agents. Data Fusion is one of the most desired processes while dealing with information processing in wireless sensor networks as the requirement of increased lifetime of network, reduced and balanced energy consumption still pre-dominates all other factors affecting the performance of sensors deployed in nondeterministic areas. In fact, data fusion serves as a solution to very common problem known as sensor hole problem. Although there exist various data fusion strategies to address the issue, but all of these suffer from one or the other drawback. Hence, this paper initially begins with a comparative overview of various data fusion strategies exploring the exploitation of software agents and later proposes a hybrid approach integrating the pros of both parallel and serial fusion approaches for achieving the same.