在文献[1]中,我们曾指出对于局部超音速点,存在一个小于1而大于零的ω_0。当松弛因子满足0<ω_0<0≤1时,改进迭代是稳定的。这个结论是错误的。错误来自在附录2中,令ω→0,而未对|B_2|~2—|A_2|~2作精密的量级分析。正确的结论是:在局部超音速点,对改进迭代找不到满足稳定条件的ω。下面证明这结论: In [1], we have pointed out that for the local supersonic point, there is ω_0 less than 1 and greater than zero. When the relaxation factor satisfies 0 <ω_0 <0≤1, the improved iteration is stable. This conclusion is wrong. The error comes from appendix 2, making ω → 0, but not the exact magnitude analysis of | B_2 | ~ 2- | A_2 | ~ 2. The correct conclusion is: at the local supersonic point, we can not find the ω that satisfies the steady-state condition for the improved iteration. Here’s the proof: