Vol.2 , issue 1
1. PaperID 013: A New Estimation Method for Calculation of Betweenness Centrality in Large Graphs (pp. 84-91) | Fulltext|
Department of computer, Islamic Azad university, Ayatollah Amoli Branch, Amol, Iran
Abstract – Calculating betweenness centrality is one way to find influential vertices of a graph and identify vertices more traversed than others. The common method takes advantage of the shortest paths between two vertices in order to find the centrality. The problem with this approach is that in many real-world applications, traversing edges and vertices does not necessarily take place in the shortest paths (news, rumors, and messages do not always pass from the shortest path to reach the target). For this purpose, random Betweenness centrality approach (or Random-Walk Betweenness) has been developed, in which the paths between source and destination are randomly selected. Random Betweenness centrality of the vertex v shows the number of traverses from vertex v in all random paths between all vertices. This algorithm has a high time complexity of O (n3) and high computational requirement; therefore, it hat a limited use for very large graphs. Motivated by this limitation, different estimation methods are presented in this research to estimate the random Betweenness centrality. Estimation methods such as random, linear and bisection have been already tested on different algorithms. In shortest path based betweenness centralities, linear estimation gives better result comparing to random and bisection estimations. Here, a novel linear and bisection estimation methods are proposed particularly for random walk betweenness and then it is shown linear method gives better results comparing to other available methods. It also requires less computational complexity comparing to bisection method. The main contribution of this work is to develop an estimation method to achieve a fair estimation of random-walk betweenness centrality using linear estimation.
Keywords: betweenness centrality, bisection estimation, centrality estimation, linear estimation, random-walk betweenness
Full Text: PDF [Academia.edu |
2. PaperID 014: New methods to improve service quality parameters TCP in Wireless Sensor Networks (pp. 92-98) | Fulltext|
Mehdi Akbari, Mohammad Mehdi Hassani
Islamic Azad University, Ayatollah Amoli Branch, Amol, Iran
Abstract– AD Hoc sensor networks are considered as a network architecture where all nodes are able to communicate with others through multi-hop routing or sending system. These AD Hoc networks become very popular because they don’t need a pre-determined structure. Using TCP/IP in sensor networks makes direct communication and connection between sensor network and output networks of TCP/IP possible. TCP indicates poor performance in the wireless medium in terms of energy, efficiency, and output. Research use extensive practical studies and simulation to understand and assess the performance of multiple wireless networks. Through investigating the performance of one TCP session in a series topology, a model of the system is presenting and proposing to analyze TCP’s performance in multiple wireless networks.
keywords— Sensor network,TCP transmission in sensor networks, Quality of Service, Congestion Control
Full Text: PDF [Academia.edu |