最优多用户检测问题研究

A Study of Optimum Multiuser Detection Problem

DS-CDMA无线通信中的最优多用户检测属于NP完备组合优化问题,启发式方法是求解这类问题的有效方法,通过分析最优多用户检测问题的适应值曲面特征,研究设计了系列低计算复杂度、接近最优多用户检测性能的启发式算法.仿真结果表明,基于演化策略的多用户检测算法能够在中等规模用户数情况下提供与最优多用户检测相当的性能,而快速迭代局域搜索算法能够以较低的计算复杂度得到比其他局域搜索算法更好的解.

Optimum multiuser detection （OMD） is an NP-complete combinatorial optimization problem in DS-CDMA wireless communication systems and heuristics are efficient methods for solving such problems. By analyzing the fitness landscape of the OMD problem, a series of multiuser detection algorithms are presented, which have lower computational complexity and good performance. The evolution strategy （ES） algorithm can achieve the performance of the OMD bound, and the fast iterated local search （FILS） method can obtain much better solution than that of other correlative algorithms with lower computational complexity.