Design And Development To Agent-oriented Simulation System Of Product Development Process

To modern enterprise, product development is a key link to improve its competitiveness. Along with product complexity increasing, there have been more and more kinds of new product development modes put forward and widely used gradually. It has become an important issue to academes and enterprises on how to make an efficient plan for product development process in this background, so as to improve product competitiveness. Different from other work flows, product development process (PDP) is a typical knowledge processing course, in which the effects of design agents and design organizations are greatly important. However, in existing researches on PDP modeling and simulating, design agents were generally treated as common resources, while the PDP model was task-oriented. In such model, the difference between design agents and resources, and design agent's behavior characteristics like initiative, autonomy and collaboration weren't taken into account, causing that the design agent's activities and organization strategies couldn't be analyzed and optimized correctly. Therefore, in this dissertation, a further study which concentrated on the PDP modeling and simulating was implemented, according to the characteristics of PDP and taking account of design agents's initiative activities during PDP, and an agent-oriented simulation system on PDP was also designed and developed. The main contents were consisted of following parts:Firstly, by analyzing the characteristics of PDP, an agent-oriented PDP modeling method was proposed, based on CAS (Complex Adaptive System) theory and MAS (Multi-agent System) modeling principle. According to the method, an agent-centered PDP integration model was built which was consisted of 5 sub-models: design agent model, task model, product information model and resource model. This integration model replaced of task-centered modeling method, and described PDP as a continuous process, in which design agents changed the status of all tasks until they were implemented under the constraints of resources and information. in this way, the integration model absolutely considered the important role of design agents in PDP, and reflected the initiative, autonomy and collaborative of design agents,setting a foundation for design and development to simulation system.Secondly, simulation flow and function frame of A-PDPS (Agent-oriented Product Development Process Simulation System) were designed based on the former simulation model, and the function of graphical modeling was also described. To develop A-PDPS, simulation flow was subdivided into 5 phases including parameter initialization, task screening, task selection, task execution, and simulation results output; function frame was separated into 4 layers, which were user layer, model layer, running layer, and data process layer; function of graphical modeling was the description to model layer in detail, as well as the realization of parameter initialization phase.Thirdly, technical implement to A-PDPS was carried out based on the former simulation flow and function frame. A-PDPS was developed by applying Visual C++. To achieve the function of the system, the integration model which included design agent and environment object was packaged respectively; design agents'characteristics in PDP like initiative, autonomy and collaboration were also packaged into different behavior functions, which were invoked frequently in simulation so as to realize the key role of designers in PDP; at the same time, the mechanism of function invoking was described in phases of graphical modeling and simulation results output. Furthermore, some product design project was simulated to testify the correction and validity of the simulation system. The simulation results showed that the A-PDPS and development method could not only describe the initiative, autonomy and collaboration behaviors of design agents, but also offer a more scientific basis for organization behaviors optimization.Lastly, simulation study was carried through on design agent's autonomy and collaboration in PDP as well as influencing factors of organization performance, while complexity of PDP in A-PDPS was testified. In this manner, simulation experiments were designed and analyzed on whether A-PDPS reflected design agents'autonomy and collaboration in PDP and how different organization structures had effect on product development cycles. According to simulation results, it showed that A-PDPS reflected design agents'characteristics, providing an efficient way to forecast and optimize PDP and improve organization performance.