A Queue-Length Based Approach to Metropolized Hamiltonians for Distributed Scheduling in Wireless Networks

Link scheduling in wireless networks for Medium Access Control (MAC) has received considerable attention over the past decade. In this work, we propose a distributed link scheduling algorithm inspired by the Ising model from statistical mechanics. This work overcomes the shortcomings of previous Ising model based algorithms by incorporating queuing and servicing dynamics for improving aggregate throughput and network latency. We develop and associate a novel Hamiltonian with a network interference graph and propose an optimization algorithm that is easily implementable in practice. Our simulations show considerable improvement in aggregate throughout and network delay compared to existing benchmarks, over different network topologies and traffic patterns.