TWO-LEVEL DECOMPOSITION METHOD FOR RESOURCEALLOCATION IN TELECOMMUNICATION NETWORKS

n this paper, we consider a two-level problem of resource allocation in a telecommunication network divided into zones. At the upper level the network manager distributes homogeneous resource shares among zones in order to maximize the total network profit, which takes into account the inner zonal payments from users and the implementation costs. This means that each zonal income calculation at a given resource share requires solution of the inner resource allocation problem. As a result, we obtain a two-level convex optimization problem involving non smooth functions whose values are calculated algorithmically. Unlike the usual convex non smooth optimization methods we suggest this problem to be solved by a Lagrangean duality method which enables us to reduce the initial problem to a sequence of hierarchical one-dimensional problems. Besides, we suggest new ways to adjust the basic problem to networks with moving nodes. We present results of computational experiments which confirm the applicability of the new method. n this paper, we consider a two-level problem of resource allocation in a telecommunication network divided into zones. At the upper level the network manager distributes homogeneous resource shares among zones in order to maximize the total network profit, which takes into account the inner zonal payments from users and the implementation costs. This means that each zonal income calculation at a given resource share requires solution of the inner resource allocation problem. As a result, we obtain a twolevel convex optimization problem involving non smooth functions whose values are calculated algorithmically. Unlike the usual convex non smooth optimization methods we suggest this problem to be solved by a Lagrangean duality method which enables us to reduce the initial problem to a sequence of hierarchical onedimensional problems. Besides, we suggest new ways to adjust the basic problem to networks with moving nodes. We present results of computational experiments which confirm the applicability of the new method. I

[1]  B. Nordstrom.  FINITE MARKOV CHAINS , 2005 .

[2]  I Konnov,et al.  Optimization based flow control in communication networks , 2004 .

[3]  Masao Fukushima,et al.  Quasi-variational inequalities, generalized Nash equilibria, and multi-leader-follower games , 2009, Comput. Manag. Sci..

[4]  John G. Kemeny,et al.  Finite Markov Chains. , 1960 .

[5]  Arthur L. Liestman,et al.  A Zonal Algorithm for Clustering An Hoc Networks , 2003, Int. J. Found. Comput. Sci..

[6]  Christopher Rose,et al.  Resource allocation for wireless networks , 1994, Proceedings of IEEE Vehicular Technology Conference (VTC).

[7]  Erkki Laitinen,et al.  Optimisation problems for control of distributed resources , 2011, Int. J. Model. Identif. Control..

[8]  P. Harker,et al.  A penalty function approach for mathematical programs with variational inequality constraints , 1991 .

[9]  Kurt Rohloff,et al.  A Hierarchical Control System for Dynamic Resource Management ∗ , 2006 .

[10]  Muhammad Aslam Noor,et al.  Quasi variational inequalities , 1988 .

[11]  Slawomir Stanczak,et al.  Resource Allocation in Wireless Networks: Theory and Algorithms , 2006, Lecture Notes in Computer Science.

[12]  Ding-Zhu Du,et al.  Ad Hoc Wireless Networking , 2004, Network Theory and Applications.

[13]  Messaoud Bounkhel,et al.  Quasi-Variational Inequalities , 2012 .

[14]  Michel Minoux,et al.  Programmation Mathématique. Théorie et Algorithmes , 2008 .

[15]  O. Nelles,et al.  An Introduction to Optimization , 1996, IEEE Antennas and Propagation Magazine.