基于免疫聚类识别的耦合功能规划方法与实例

Coupling function planning & practical example based on immune clustering recognition approach

公理设计是以独立公理为基础的,并未提供分析和处理具有功能耦合关系的耦合设计手段。针对这一问题,提出了一种耦合功能规划的新方法。该方法首先利用免疫算法对功能耦合关系进行聚类识别,然后对功能耦合程度加以量化处理,最后通过解耦和割裂操作对耦合功能进行重组,从而确定各功能的实现顺序。其中耦合关系的聚类识别旨在为解耦操作提供信息支持,以避免耦合信息的不必要丢失。结合卧式壳管式冷凝器产品的设计实例,对该方法及其有效性进行了说明和验证,还对免疫聚类识别算法中的有关参数进行了讨论,并围绕算法性能的优劣和算法结果对解耦的影响作用,分别将其与遗传聚类识别算法和基于图论的聚类识别方法进行了比较。

Axiomatic Design is based on function independence axiom whereas it couldn＇t deal with coupled design involving functional coupling relationships. In order to solve the problem, a new coupling function planning approach was proposed. Firstly, functional coupling relationship was clustered and recognized by Immune Clustering Recognition Algorithm （ICRA）. Then, functional coupling degree was measured. Finally, the coupling functions were rearranged via decoupling and tearing operations in order to define realization sequences of various coupling functions. The clustering recognition of functional coupling relationship was aimed to provide support for decoupling and to avoid unnecessary loss of coupling information. A practical example of horizontal shell and tube condenser was studied to verify the effectiveness of the proposed approach. Certain parameters in ICRA were also discussed. Furthermore, the performance of ICRA was compared with that of genetic clustering recognition algorithm, and the influence of the result of clustering recognition on decoupling was also compared with that of graph- based clustering recognition method.