No abstract

No abstract

-The wired mode of communication will no longer serve the purpose of our present scenario network where the expectations in communications system have been skyrocketed. Thus, wireless network with ad hoc nature will suit the present and forthcoming real world applications. This ad hoc wireless network will bind the entire globe and will reign the communication system tomorrow. But, ad hoc network has been limited by the growth factor. In order to address this issue, the clustering mechanism has been incorporated into this ad hoc network to get better results in terms of the network functionality. The ad hoc network efficiency has been improved when the clusters have been devised properly. There are various parameters which identify the efficacy level of clusters in networks. Existing cluster formation algorithms give more weight to any one of the parameters (degree, distance, energy, mobility) when selecting the cluster head of the cluster. This selection of parameters should be corresponding to the application to which the clusters are devised and cluster heads are selected. There is a desperate need for identifying the appropriate cluster head based on the importance given to parameters at various scenarios. Thus, this paper proposes NNPAC (Neural Network based Partitioning Around Cluster head), a generic method of cluster head election to fit for different applications. Implementation of this work has been carried out with the help of MATLAB simulator.

No abstract

Ad hoc networks are formed by intermediate nodes which agree to relay traffic. The link between nodes is broken when a node rejects to relay traffic. Various parameters like depreciation in the energy of a node, distance between nodes and mobility of the nodes play a vital role in determining the node's rejection to relay traffic. The objective of this paper is to propose a novel model that identifies the cooperative nodes forming stable routes at the route discovery phase. The weight factor of the different parameters decides the varied type of networks where the proposed model can be applied. Hence, an Artificial Neural Network based non-deterministic generic predictive model is proposed to identify the varied types of networks based on the weight factor. This study has been substantiated by simulation using OMNET++ simulator. We are sure that this paper will give a better solution to identify cooperative nodes thereby improving the performance of the network. General TermsArtificial Neural Network, non-deterministic generic predictive model.