虽然光学透镜象差的自动平衡已取得非常显著的成绩,达到了实用化的地步,但它严重地依赖于初始结构的给定;初始解选得不好,往往就会失败。可是对初始解的自动创成却没有成功的报道。其主要原因,首先是所用的最优化方法大都只能得到局部极值,而初始解的创成乃是一个在大范围中求带约束的总极值问题,所以沿用通常的最优化方法不能得到良好的效果;其次,由于大范围的搜索经常发生违反边界条 Although the optical lens aberration self-balancing has achieved remarkable success and reached its practical point, it depends heavily on the given initial structure. If the initial solution is poorly selected, it often fails. However, the automatic creation of initial solutions has not been reported successfully. The main reason is that most of the optimization methods used can only get the local extremum. The initial solution is a total extremum problem with constraints in a large area, so it can not be obtained by the usual optimization methods Good results; secondly, violations of the boundary bar often occur due to a wide range of searches