Fundamental Research for Optimization of Lot Partition of Residential Blocks

提出了一种估算地块边界线的统计方法，用于估算地块形状。更具体地说，设计了用于估计最可能的边界线的编程代码，其将暂时调整给定的边界线，以使所产生的地块形状符合地块信息，例如临街长度和地块面积。边界线的预测可以表示为两个目标优化函数，一个是似然度最大化，另一个是批次形状变化最小化。引入模糊决策支持理论的概念，解决了这一问题。将该方法应用于东京的330个样本地块，正确估计了约85%的地块，证明了该方法的有效性。通过地块之间的形状相似度矩阵，给出了地块中典型地块的提取方法。该方法被应用于东京Setagaya Ward的20个街区。无论区块的形状如何，类似矩形的地块往往被选为典型地块。这是受这样一个事实的影响，即在现实中，矩形地块在选择地块时往往是首选的。在矩形块中，有一条边的地块，其长度大约等于块的较短边的长度，如典型的地块。此外，两个前缘相差2倍且深度相等的地块往往被选为典型地块。这些结果很好地反映了该区块的细分历史。如上所述，我可以开发基本技术来最佳地划分区块，我还可以开发一种方法来为区块找到典型的地块。在此基础上，以购房和租房家庭数据库为基础，分析了住宅市场的需求结构，分析了消费者对居住环境的评价。结果表明，不同时间类型的家庭对室内质量和室外质量的重视程度不同。

A statistical method for estimating the boundary lines of lots are developed for estimating the lot shape. More specifically, programming codes are devised for estimating the most likely boundary lines, which will adjust tentatively given boundary lines so that the resulting lot shape conforms to lot information such as length of frontage and lot area. The prediction of boundary lines can be formulated as double purpose optimization functions consisting of the maximization of likelihood and the minimization of change in lot shape. This problem is solved introducing a concept of fuzzy decision support theory. The method is applied to 330 lots in Tokyo as sample lots, and about 85% of the lots are properly estimated, showing the validity of the developed method. A method for deriving typical lots in a block is also developed through the shape similarity matrix among lots. The method is applied to 20 blocks in Setagaya Ward in Tokyo. Rectangular like lots tend to be chosen as typical lots irrespective of the shape of blocks. This is influenced by the fact that rectangular lots are often preferred in the choice of lots in reality. In rectangular shaped blocks, a lot having an edge, whose length is approximately equal to the length of shorter edge of the block, as a typical lot. Moreover, two lots, whose frontages differ in the factor of 2 and whose depths are equal, tend to be chosen as typical lots. Those results expresses well the history of subdivision of the block. As described above, I could develop fundamental technique to partition blocks optimally, and I could also develop a method to find typical lots for the block. Furthermore, demand structure in housing market is analyzed based on the database of households who bought or rented houses, in order to analyze how residential environment are valued by consumers. The results suggests that there are hour types of households, who differ in the priority of valuing in-door quality and out-door quality.