Parallel and distributed computation: numerical methods
Parallel and distributed computation: numerical methods
On the convergence of the coordinate descent method for convex differentiable minimization
Journal of Optimization Theory and Applications
Fair end-to-end window-based congestion control
IEEE/ACM Transactions on Networking (TON)
Convex Optimization
A scalable model for channel access protocols in multihop ad hoc networks
Proceedings of the 10th annual international conference on Mobile computing and networking
Optimization of Polynomials on Compact Semialgebraic Sets
SIAM Journal on Optimization
Geometric programming for communication systems
Communications and Information Theory
IEEE/ACM Transactions on Networking (TON)
Utility-optimal random access: reduced complexity, fast convergence, and robust performance
IEEE Transactions on Wireless Communications
The capacity of wireless networks
IEEE Transactions on Information Theory
Stability and delay of finite-user slotted ALOHA with multipacket reception
IEEE Transactions on Information Theory
IEEE Journal on Selected Areas in Communications
A Game-Theoretic Framework for Medium Access Control
IEEE Journal on Selected Areas in Communications
Hi-index | 0.01 |
Random access protocols, such as Aloha, are commonly modeled in wireless ad-hoc networks by using the protocol model. However, it is well-known that the protocol model is not accurate and particularly it cannot account for aggregate interference from multiple interference sources. In this paper, we use the more accurate physical model, which is based on the signal-to-interference-plus-noise-ratio (SINR), to study optimization-based design in wireless random access systems, where the optimization variables are the transmission probabilities of the users. We focus on throughput maximization, fair resource allocation, and network utility maximization, and show that they entail non-convex optimization problems if the physical model is adopted. We propose two schemes to solve these problems. The first design is centralized and leads to the global optimal solution using a sum-of-squares technique. However, due to its complexity, this approach is only applicable to small-scale networks. The second design is distributed and leads to a close-to-optimal solution using the coordinate ascent method. This approach is applicable to medium-size and large-scale networks. Based on various simulations, we show that it is highly preferable to use the physical model for optimization-based random access design. In this regard, even a sub-optimal design based on the physical model can achieve a significantly better performance than an optimal design based on the inaccurate protocol model.