本文提出了一个电路中心设计(DC)的新算法。在正态分布、同一方差的假设下,借助随机逼近理论讨论了它的收敛性。因其只要求对电路进行Monte Carlo分析(MCA),不用计算灵敏度,故适用于线性和非线性电路,易于在计算机上实现。算例表明,本算法收敛快,计算量少,对非凸、非单连通可行域均有效。 This paper presents a new algorithm for circuit center design (DC). Under the assumption of normal distribution and same variance, its convergence is discussed by means of stochastic approximation theory. Because it only requires Monte Carlo analysis (MCA) of the circuit, without calculating the sensitivity, it is suitable for linear and nonlinear circuits, easy to implement on a computer. The example shows that the proposed algorithm converges quickly and has less computational complexity and is effective for non-convex and non-connected feasible regions.