无限维 Hilbert 空间中, 解凸可行问题的平行投影算法通常是弱收敛的. 本文对一般的平行投影算法进行改进, 设计了一种解凸可行问题的具有强收敛性的新算法. 该算法主要是在原有算法基础上引入了一个参数序列, 在参数序列满足一定的控制条件下保证了算法的强收敛性. 为了简单证明算法的强收敛性, 我们构建了一个新的积空间, 然后把原空间的这种改进平行投影算法转换为积空间中的交替投影算法. 这样，改进的平行投影算法的强收敛性就可以通过交替投影算法的收敛性证明得到.

It is well known that the classical parallel projection algorithm for convex feasi- bility problem in Hilbert space is weak convergent. In this paper, a modification of parallel projection algorithm is presented by introducing a parameter sequence for solving the con- vex feasibility problem. To prove the strong convergence in a simple way, we introduce a product space. Then, we transmit the modified parallel algorithm in the original space to a aternating one in the product space. Thus, the strong convergence of the modified parallel projection algorithm is derived from the alternating one under some parametric controlling conditions.