Abstract: Software size estimation is the key of entire software program project，and the exact estimation immediately relates to the success of the project.Traditional Function Point Analysis（FPA） has the deficiency of complexity grade analysis.In order to deal with the problem，an improved method for computing complexity weight is proposed.The method analyzes the complexity weight of function component by the fuzzy theory.Firstly，fuzzy logic system for analyzing weight is established.The inputs of fuzzy system are Data Element Types（DET） and Record Element Types（RET），and the output is the component’s complexity weight.Then，according to the relationship between complexity and function point，the fuzzy rules are set.By using these fuzzy rules，the complexity weight can be computed.The results of research show that the method can eliminate discontinuity among the different grades and enhance the accuracy of function point estimation.

Key words: function point analysis, software size estimation, fuzzy rules, complexity weight

摘要： 规模度量是软件项目管理的关键，其度量的准确性直接关系到软件项目的成败。针对传统FPA方法中复杂度等级划分不连续性的问题，提出一种改进的复杂度权值计算方法。该方法利用模糊理论分析功能要素的复杂度，首先以DET和RET作为输入变量，以复杂度权值作为输出变量，建立模糊推理系统；然后根据FPA中复杂度和功能点数量的转换关系，设置模糊推理规则，并利用该规则确定复杂度权值。研究结果表明，这种模糊推理的方法可以消除不同复杂度等级之间的断层，从而使软件功能点的估算结果更加准确。

关键词: 功能点分析, 软件规模估算, 模糊规则, 复杂性权值