针对凸不等式系统提出了一种具有可变权参数和下松弛因子的平行不完全投影算法,经过有限次的投影算法得到不完全投影点,并且通过构造一个新的积空间,将欧氏空间中的平行不完全投影算法转化为新空间中的半序列不完全投影算法,使得平行的不完全投影法的收敛性由半序列不完全投影算法的收敛性证明直接得到,在一定程度上简化了平行不完全投影算法的收敛性证明.
A parallel incomplete projection method with variable weights and under relaxation coefficient was proposed for the system of convex inequalities. Then, the parallel incomplete projection method was transformed in a given Euclidean space into a semi-sequential incomplete projection method in a product space by constructing a new space. Convergent results for the parallel incomplete projection methods may be obtained from the semi-sequential ones, which can simplify the proof of the convergence.