基于复杂系统视角的服务型制造信息系统鲁棒运作机理与度量研究

Various participants and random requirements render the service-oriented manufacturing information system to increasing complexity, dynamic, and uncertainty. Given the realities that conventional systems have inadequate ability to deal with it and that insufficient work has been done in literature, research project on robustness measurement and mechanism of robust operation for a service-oriented manufacturing information system is proposed to study the comprehensive anti-jamming capability of information systems, which mainly adopts the method of synergetics combining with the stability theory, software engineering and multi-agent simulation. The project establishes a set of indexes which measure robustness and models of functional structure and logical structure for typical systems by sufficient investigation and analysis. And then a dynamic evolution model for robust operation is constructed. The research finds out the modes how order parameters act on system parameters and criteria of robustness measurement to reveal the mechanism of robust operation. Combining uncertainty of customers, resources and tasks, the models proposed are verified and improved by cases study and multi-agent simulation. The research is an innovative try in solving problems of robust operation in information systems through methods of Science of Complexity. It is expected to settle bottleneck problems for service-oriented information systems with weak robustness, to be the theoretical basis to control and optimize weakly robust systems. Furthermore, the project provides theoretical guides to management practices such as robustness prediction, decision making, and is of great significance to the transformation and upgrading of service-oriented manufacturing industry.

参与主体多样化和需求随机性变化使得服务型制造信息系统呈现出更为复杂的动态性和不确定性，鉴于常规系统无法应对的现实问题与理论上研究的不足，以协同学为主，结合稳定性理论、软件工程和多Agent仿真等技术方法，从复杂系统动态演化视角提出服务型制造信息系统鲁棒运作机理与度量研究课题。基于充分的调研分析，构建典型系统的功能结构模型、逻辑模型及鲁棒性度量指标体系，建立系统鲁棒运作动态演化模型。通过模型研究序参量对系统变量的役使方式及系统鲁棒性度量判据，揭示系统的鲁棒运作机理，形成系统鲁棒性度量方法；结合客户、资源和任务等实际不确定情形，利用多Agent仿真及案例验证和完善系统模型体系。该研究是用复杂系统方法解决服务型制造信息系统鲁棒运作瓶颈问题的创新性尝试，有望成为系统鲁棒控制与优化的基础理论，对信息系统鲁棒运作设计与规划、鲁棒性预测与决策等管理实践提供理论指导，对制造业服务化转型升级具有重要意义。

与传统制造信息系统相比，服务型制造信息系统是多组织跨区域多层次的复杂网络，具有开放性、参与主体多样性、复杂性和动态性等新特性，不确定性因素影响显著增加，造成系统的失效风险和破坏程度更大。因此，服务型制造信息系统的鲁棒性控制尤为关键。本项目以协同学为主，结合稳定性理论、软件工程和多Agent仿真等技术方法，从复杂系统动态演化视角提出服务型制造信息系统鲁棒运作机理与度量，分别建立系统鲁棒运作动态演化模型，形成系统鲁棒性度量方法，完善服务型制造信息系统模型体系。相关研究成果有利于丰富和拓展这些理论体系，促进我国制造业转型升级提供重要理论支撑。