Title: Improved neural approach in maximising reliability for increasing networks
Authors: Baijnath Kaushik; Navdeep Kaur; Amit Kumar Kohli
P.T.U. Jalandhar, PTU-Kapurthala Highway, Near Puspa Gujral Science City, Kapurthala-144601, India
Computer Science Department, Chandigarh Engineering College, Landran, Mohali, Punjab-140307, India
Electronics & Communication Engineering Department, Thapar University, Patiala, Punjab, 147004, India
Abstract: A method is presented to maximise reliability for increasing networks. A neural approach is combined with reliability values of each link obtained from minimal cuts in increasing network. The method simply evaluates minimal cuts from highly increasing networks and a two-dimensional combinatorial spectrum is obtained from an approximation formula for assigning reliability values for each link. These reliability values will be used in a neural approach as the upper-bound on reliability for improving reliability. An increasing network is considered with random failure in links and nodes. Evaluating minimal cuts in increasing networks requires significant computational effort, but, when approximated, computational time reduces significantly. The result shows significant improvement in reliability for increasing networks design when an approximated combinatorial spectrum is used as input to the neural networks. The approach reduces significantly the computational effort for reliability calculation.
Keywords: adaptive gradient neural networks; combinatorial spectrum; minimal cuts; minimal path; reliability; increasing networks; networks design.
Int. J. of Computational Science and Engineering, 2015 Vol.11, No.2, pp.176 - 185
Available online: 22 Sep 2015