Abstract:

In order to enhance the global exploration ability of the binary artificial bee colony algorithm, a binary artificial bee colony algorithm based on estimation of distribution is proposed. The proposed algorithm is applied on the optimal multiuser detection technique, and the multiuser detection scheme based on the algorithm is designed. The algorithm directly adopts the multidimensional neighborhood search strategy in the discrete domain, thus quickening the convergence speed and avoiding the conversion from continuous domain to discrete domain. Meanwhile, it makes use of the global statistics information obtained by estimation of distribution to generate the candidate solutions. The simulation results show that, compared with the conventional detectors, the convergence speed of the designed detector is faster, the bit error rate and the nearfar resistance ability is significantly improved.

Key words: optimum multi-user detection;artificial bee colony algorithm;binary artificial bee colony algorithm;estimation of distribution algorithm