Chunks and power allocation in OFDMA systems based on transient chaotic neural network

Abstract

The issue of dynamic resource allocation in OFDMA downlink system is investigated based on Rate Adaptive (RA) Principle. A two-step adaptive resource allocation algorithm is developed so as to reduce computational complexity. The proposed algorithm consists of chunks (a set of contiguous subcarriers) and power allocation. Firstly, in order to balance the total throughput of system and fairness among users, the number of chunks is allocated proportionally to users according to their service priorities. And under the assumption that power is allocated averagely to each subcarrier, chunks are allocated to each user by transient chaotic neural network (TCNN), which can find the optimal solution of predefined optimization problem through its rich neurodynamics. Secondly, power is reallocated to chunks by classical water-filling algorithm in order to further improve the performance of system. Simulation results show that the proposed algorithm can improve the throughput of system efficiently compared with the chunk-based resource allocation (CRA) and Average power allocation (APA) algorithm.