Improved Estimation of Distribution Algorithm to Solve Multiuser Detection

In order to overcome the shortcoming of premature for estimation of distribution algorithm, a diversity enhancement estimation of distribution algorithm was proposed. Meanwhile, the improved algorithm was used to optimize the problem of multiuser detection. The diversity determination and diversity enhancement operation were introduced into the improved algorithm. The independent individual density was used to evaluate population diversity. When the independent individual density is lower than the diversity threshold, the diversity enhancement operation was performed by random variation, which can avoid premature convergence. At the same time, in order to prevent outstanding individual being consumed by diversity enhancement, the elite retention strategy was added to the sampling procedure. The simulation results show that the detection proposed in this paper has faster convergence speed, can effectively avoid premature convergence and successfully find the global optimal detection vector. It has the same anti-multiple access interference performance and near-far resistance ability performance with the optimal multiuser detection.