Premium
Levenberg‐Marquardt optimization method for coverage and connectivity control in backbone‐based wireless networks
Author(s) -
Khan Khalid,
Lobiyal D. K.
Publication year - 2017
Publication title -
international journal of communication systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.344
H-Index - 49
eISSN - 1099-1131
pISSN - 1074-5351
DOI - 10.1002/dac.3296
Subject(s) - levenberg–marquardt algorithm , computer science , gradient descent , method of steepest descent , mathematical optimization , convergence (economics) , newton's method , rate of convergence , optimization problem , cauchy distribution , wireless , wireless network , algorithm , artificial neural network , mathematics , telecommunications , nonlinear system , channel (broadcasting) , artificial intelligence , statistics , physics , quantum mechanics , economics , economic growth
Summary One of the main challenge in designing wireless networks is to ensure optimal coverage and connectivity, which can be achieved by optimally repositioning nodes in backbone network such that the total energy requirement is minimized. Therefore, an efficient optimization algorithm that converges at faster rate while minimizing the cost function (total power) is required. In this paper, we propose the use of Levenberg‐Marquardt optimization algorithm to achieve assured coverage and connectivity control in backbone‐based wireless networks. The Levenberg‐Marquardt method combines the advantages of Cauchy steepest descent method and Newton‐Raphson method and is expected to achieve faster convergence rate irrespective of initial conditions, which is the main motivation of this study. It relies on that when current solution is far from the optimal solution, then steepest descent method will be used, and as current solution approaches to optimal solution, then gradually switched to Newton‐Raphson method to find the optimal solution. Extensive simulations using M A T L A B R 2020 a has been performed to demonstrate the effectiveness of proposed method by measuring its performance in the number of iterations, elapsed time, and power requirement to maintain proper coverage and connectivity. The performance is compared with Cauchy steepest descent method as well as Newton‐Raphson method.