自适应微调扰动和声搜索算法几何约束求解研究*

几何约束求解的方法关系到特征造型系统的性能，为提高几何约束求解的速度，将和声搜索算法应用于几何约束求解中。通过优先选择较小的和声库，利用最好解的评价值确定微调扰动的幅度，并将其嵌入到拉斯维加斯算法中，提高了和声搜索算法的性能。实验结果表明，改进的和声算法具有自适应性，能有效克服局部收敛问题，提高了求解速度。

Research on adaptive fine-tuning disturbance harmony search algorithm for geometric constraints solving

Performance of feature modeling system relates to methods of geometric constraint solving. To improve the speed of geometric constraint solving, this paper applied harmony search algorithm to the geometric constraint solving. In order to increase performance of harmony search algorithm, selected a smaller harmony memory size preferentially, employed the assessment value of the best solution to determine the range of the fine-tuning disturbance, and embedded it into the Las Vegas algorithm. Experimental results show that the improved algorithm is adaptive, and that can overcome the problem of local convergence effectively, while improving the speed of the solution.