تجزیه و تحلیل مدل سازی و عملکرد شبکه های روند توسعه محصول

Product development process (PDP) is a complex system, in which design activities and design resources are connected to each other according to specific rules. Considering the elements of PDP (complex forms of design activities, interdisciplinary collaboration and multiform resources) and relationships between them, a product development process network (PDPN) is proposed and its performance is further analyzed and optimized. In order to establish the PDPN, design activities and resources used in the PDP are mapped into the nodes of the network. And the relationships among the nodes are mapped into the edges. Then, a level-based evaluation model is proposed to illustrate the forming procedure of establishing the PDPN. Then, some typical topological and physical characteristics of the PDPN are defined and analyzed based on the complex network theory. In order to balance the resource load, the PDPN is divided into different communities, wherein a Girvan–Newman (GN) algorithm together with improved measure criterion is applied. Finally, the design process of a steam turbine rotor is chosen as an example to illustrate the feasibility and availability of the proposed method.

Product development process (PDP) is considered to be a specific process of converting customers' need into a technical and commercial solution (Whitney, 1990). Each PDP is unique, but they may share some common features or elements. Therefore, it is important to model and simulate PDP with the purpose of guiding the management of future product development process (Smith and Morrow, 1999). Accordingly, an increasing amount of attentions has been paid to the construction of PDP models. Many different types of models have been found in literature, such as PERT/CPM, SADT/IDEF0, Petri net, signal flow graph, design roadmap, DSM, etc. Based on these models, PDP performances such as iteration, time, efficiency, etc., have been discussed from different perspectives in order to analyze PDP and improve its efficiency. Actually, all this research has made a great contribution to the further study of PDP. PDP can be divided into several phases, and each phase can be decomposed into a large number of design activities. Therefore, PDP can be considered as a large, complicated and unstructured network. Actually, not only the design activities but also other elements such as human resource, hardware resource, software resource, knowledge resource, etc., are involved in PDP which are neglected in the traditional research and models. As a result, these models cannot reflect the real situation of PDP and their solution turn out to be incomplete. For example, only design activities or human resources are mentioned in the Petri net. This weakness results in an information loss of PDP, which means only partial information can be obtained in this process. Performance analysis and optimization is another important aspect of PDP research. Traditional research is incomplete in this aspect because of the neglect of resources and the relationships among them. In fact, the resources and relationships play important roles in PDP from the management viewpoint. Taking resources and relationships into consideration, PDP can be described as a more complex network than the conventional ones mentioned above. The complexity of the network structure makes it is hard to analyze and optimize its performance by using the traditional models. Therefore, some new methodologies dealing with this problem from the complex system viewpoint are needed, such as the complex network theory. To this end, a new model for PDP called the product development process network (PDPN) is proposed in this paper, which is constructed based on the complex network theory. First, resources involved in the PDPN are investigated from previous studies and experience, which can be classified into three groups, namely human resource, tool resource and knowledge resource. According to the classification, design activities and resources are abstracted as the nodes of the network. Different relationships among the design activities and resources are abstracted as the edges. The generation rule and formation description of the PDPN are defined based on relation algebra. Then, several typical topological and physical characteristic indicators are defined in order to present the PDPN performance. Furthermore, with the purpose of optimizing the performance of the PDPN, the PDPN is divided into communities by using the original GN (Girvan–Newman) algorithm with the physical-characteristic-based criterion. Finally, on the basis of a case study, we discuss how the proposed method simulates the product development process, analyze and optimize the performance of the PDPN. The rest of this paper is organized as follows. After a brief review about the related fields of PDP and its applications to indicate the background and motivation of this research in Section 2, the construction process of PDPN, such as the elements, formation description and generation rules are put forward in Section 3. And then, Section 4 shows the topological and physical characteristics of PDPN which are used to describe the performance of the network. In Section 5, an improved GN algorithm is proposed for the community detecting with the purpose of PDPN performance optimization. In Section 6, the design process of a steam turbine rotor is chosen as an example to illustrate the feasibility and availability of the proposed method. Finally, some analysis and discussions are drawn in Section 7 and the conclusions and future works are summarized in Section 8.

This paper presents a model named product development process network (PDPN) for PDP modeling, performance analyzing and optimization. In order to support this model, several models and methods are introduced in this paper. The implementation of PDPN includes three steps as follows: Design nodes and resources nodes are abstracted from the design activities and resources firstly in step 1, which shows the basic elements of the model. Then, a level-based evolution model is introduced, and the formalized description of PDPN is shown by using the relation algebra. Secondly, some typical topological and physical characteristics are defined in step 2, and the mathematic formulations are given after the definitions. All these characteristics can offer the basic information for PDPN optimization. Then, a physical-characteristic-based criterion is proposed to improve the original GN algorithm. Taking physical characteristics into consideration, this new criterion can make a new balance between topological and physical characteristics for the community detecting. Finally, the design process of a steam turbine rotor is studied as an example to demonstrate the feasibility and availability of the proposed model. As for further research, the following can be considered. First, in order to reflect real situation of PDP, the quantification of the weight values among different nodes should be conducted in the near future. This value will be closely related with the specific physical meaning of PDPN. What’s more, more physical characteristics should be defined for the purpose of analyzing and optimizing performance of PDPN.