D-优化算法实现几何约束问题的重组

摘要

当前对几何约束求解普遍使用建设性方法,按照先分解后重组的方式进行求解.但多数情况下这些方法产生非线性方程数量多,计算量大,解决问题较为复杂.此处提出D-优化算法通过将欠约束子问题分离的方法,利用自定义的分离规则,将问题分解为若干单位问题.单位问题规模较小,求解难度较低.实验证明,该优化方法可有效降低问题求解的复杂程度.%The geometric constraint solvier always uses constructive approach currentely,and the model of this approach is decomposition and restruction. However,in many times,this approach will generate lots of nonlinear equations. So it's too complex. In this paper we try to use D-optimization algorithm to decompose the under-constrained problem from a GCP, and use the decomposition rules which are defined by myself to transform it into several unit problems. The unit problem can be solved much more easier. The experiments proved that this algorithm can bring down the complexity of GCS effectively.