Application System Design based on Constraint Programming Approach

本文的主要研究成果如下：1.约束规划作者讨论了一种适用于文书工作的新型约束规划方法。办公室中的业务规则可以很容易地转换为约束编程风格。约束编程具有较高的模块化和可维护性。这种方法的一个主要问题是在面向对象程序设计方法中对分层数据对象的适用性。要解决这个问题，还需要进一步的研究。人们认为，XML可能是解决所提议方法的问题的一个很好的解决方案。工作流设计和投资措施效率低下的办公业务工作流导致了约束规划方法中的硬编程任务。因此，作者提出了一种使用作业成本计算或作业基础管理(ABC/ABM)来优化工作流系统的新方法。通过IEICE(The Institute of Electronics，Information，and Communication Engineering)开发的实际应用系统对该方法进行了实验验证，实验结果表明该方法是实用有效的。约束满足与隐私保护我们明确了信息整合侵犯隐私权问题等同于一种约束满足问题。从这些观点出发，对一些通过互联网侵犯隐私权的案件进行了讨论。约束满足观可以澄清和分析侵犯隐私权的过程。然而，进一步的研究还需要采用概率约束满足方法。

The major research results are summarized as the followings :1. Constraint ProgrammingThe authors discussed a new type of constraint programming approach for clerical works. Business rules in offices can be easily converted into the constraint programming style. Also, the constraint programming has high modularity and maintenance-ability. A major problem of this approach is applicability for hierarchical data objects in object oriented programming approach. Further research was required to solve this problem. It is considered that XML may be a good solution to resolve the problem of the proposed approach.2. Workflow Design and Investment MeasurementInefficient workflow of clerical business results in hard-programming task in the constraint programming approach. Thus, the authors proposed a new approach to optimize workflow systems using ABC/ABM (Activity-based Costing or Activity-based Management). The proposed method was evaluated experimentally by a real application system developed by IEICE (The Institute of Electronics, Information, and Communication Engineers) and the experimental result shows that the proposed method is practical and effective.3. Constraint Satisfaction and Privacy ProtectionWe clarified that "privacy infringement by information integration" problem is equal to a kind of constraint satisfaction problem. From the viewpoints, some privacy infringement cases through the Internet were discussed. The "constraint satisfaction view" can clarify and analyze the privacy infringement process. However, probabilistic constraint satisfaction approach should be required for the further research.