Simulated Annealing based multiuser detection for synchronous SDMA system

In this treatise, a novel simulated annealing (SA) based multi-user detection (MUD) is proposed in synchronous space division multiple access (SDMA) system. SA MUD modifies experiential cooling schedule (CS) of traditional SA algorithm according to its use in MUD. Moreover, in order to ensure sufficient diversity acquired in the whole Markov chain and to prevent from being trapped at local optima, uniform mutation (UM) based trial vector generation scheme is brought forward. In addition, the optimal solution recording scheme is also invoked in case of being lost during cooling process. Simulation results illustrate that in comparison with genetic algorithm (GA) MUD in the same simulation conditions, without turbo processing and soft-information, SA MUD proposed in this paper performs better, approaching the performance of maximum likelihood (ML) MUD and imposes lower complexity.