Cross-Layer Optimization Based on Partial Local Knowledge (Special Paper)

This paper considers decentralized cross-layer optimization for multichannel random access with partial-local knowledge. In the network scenario considered, users are not necessarily within the transmission ranges of all other users, and each user may choose to send packets to or receive packets from different users simultaneously. A criterion for cross-layer optimization is provided leading to a decentralized optimization for multichannel random access (DOMRA) scheme. DOMRA can be applied to different types of wireless networks, such as wireless sensor networks and mobile ad hoc networks, to improve quality of service. Furthermore, DOMRA can be applied to optimize downlink transmissions in cellular networks for proposed cochannel interference avoidance (CIA) medium access control (MAC). CIA-MAC requires low signalling overhead and minor changes to existing mobile systems. The conditions for triggering CIA-MAC are investigated and a simple trigger mechanism is obtained. Simulation results confirm the significant performance improvement by the proposed schemes.