随机树生成算法设计及分析

传统的随机树生成算法对于大规模树状结构数据集的可视化算法测试有很大的局限性。在经典随机树生成算法-GROW算法的基础上提出了两种新的随机树生成算法,实验表明它们均可以生成接近于期望规模的随机树,但在平衡性、时间复杂度、最终生成树的规模、宽度与深度的随机性方面表现各异。同时对这些特性与相关参数的关系作了探讨,实验结果分析表明参数因子的选取直接影响着生成随机树的最终表现,对于随机算法的选取有着很好的指导性作用。

Design and Analysis of Random Tree Generation Algorithms

The traditional random tree generation algorithms aren＇t suitable for generating large-scale hierarchical structures used for testing visualization layout algorithm. In this paper, we propose two new random tree generation algorithms based on the classical random tree generation algorithm-GROW, the experiment shows that they can generate the random trees which are close to the expected size, but they behave differently with the balance and several other statistical properties such as the mean time-complexity, size, depth and branches. We also discuss the relationship of these characteristics and related parameters.