Methodology and Computing in Applied Probability ( IF 0.9 ) Pub Date : 2021-02-07 , DOI: 10.1007/s11009-021-09849-7 Weiguang Peng , NingNing Peng , Kazuyuki Tanaka
Okisaka et al. (2017) investigated the eigen-distribution for multi-branching trees weighted with (a,b) on correlated distributions, which is a weak version of Saks and Wigderson’s (1986) weighted trees. In the present work, we concentrate on the studies of eigen-distribution for multi-branching weighted trees on independent distributions. In particular, we generalize our previous results in Peng et al. (Inform Process Lett 125:41–45, 2017) to weighted trees where the cost of querying each leaf is associated with the leaf and its Boolean value. For a multi-branching weighted tree, we define a directional algorithm and show it is optimal among all the depth-first algorithms with respect to the given independent distribution. For some balanced multi-branching trees weighted with (a,b) on the assumption 0 < r < 1 (r is the probability that the root has value 0), we further prove that if an independent distribution d achieves the distributional complexity, then d turns out to be an independent and identical distribution.
中文翻译:
独立分布下多分支加权树的本征分布
冲坂等。(2017)研究了在相关分布上用(a,b)加权的多分支树的特征分布,这是Saks和Wigderson(1986)加权树的弱版本。在当前的工作中,我们专注于在独立分布上对多分支加权树的特征分布进行研究。特别是,我们在Peng等人中概括了我们以前的结果。(通知Process Lett 125:41–45,2017年)到加权树,其中查询每片叶子的成本与叶子及其布尔值相关。对于多分支加权树,我们定义了一种方向算法,并表明对于给定的独立分布,在所有深度优先算法中它是最佳的。对于一些加权的平衡多分支树a,b)假设0 < r <1(r是根的值为0的概率),我们进一步证明,如果独立分布d达到分布复杂度,则d成为独立且相同的分布。