由于 Verdu 提出的最优多用户检测方法采用Viterbi算法 ,它的计算复杂度与用户数成指数关系 ,无法在实际中应用 ,因此提出了一种采用半正定规划 (positive se-mi- definite programming,SDP)的新的准最优多用户检测方法。这种方法将最优多用户检测方法转化成一个二次规划(quadratic programm ing,QP)的求解问题 ,并通过半正定规划松弛来解决这个二次规划问题。由于这种方法采用的半正定规划问题求解方法具有多项式程度的复杂性 ,所以这种新的 SDP多用户检测方法也具有多项式程度的复杂性。仿真表明 :这种新的 SDP多用户检测器的性能十分逼近最优多用户检测器的性能 Because Verdu proposed the best multi-user detection method using Viterbi algorithm, its computational complexity and the number of users in exponential relationship, can not be applied in practice, so a positive semi-definite programming (positive se-mi-definite programming, SDP), a new quasi-optimal multi-user detection method. This method transforms the optimal multi-user detection method into a quadratic program (QP) solving problem and solves this quadratic programming problem by semi-definite programming slack. Because of the polynomial degree of complexity of the solution to the semi-definite programming problem adopted by this method, the new SDP multi-user detection method also has the complexity of polynomial degree. Simulation shows that the performance of this new SDP multiuser detector is very close to the performance of the optimal multiuser detector