Robustness Analysis And Control Of Smart Manufacturing Information Systems

Smart manufacturing,an innovative megatrend in worldwide manufacturing industry,is and will continue for a long time to be the main direction and pivotal sally port for the transformation and upgrading of Chinese manufacturing industry.The smart manufacturing information system(SMIS)emerges as a vital informatization means supporting the operations of smart manufacturing.Its robustness against internal and external perturbations is not only the crucial premise and guarantee to steadily propel smart manufacturing,but exists as one of the key scientific problems that must be solved in manufacturing intelligentization.To enhance the ability of information systems to resist or adapt to environmental changes,industry tends to use engineering approaches,such as adopting new technologies,upgrading or developing functional modules,strengthening infrastructure.However,due to the development philosophy of "emphasizing technologies and neglecting management",such engineering approaches restrain systems from exhibiting expected performance.Existing literature often views information systems as technical systems,and mainly employs technical solutions to improve the robustness of local systems,such as software,operating systems and computer systems.It ignores the fact that information systems are the syntheses composed of both organizational systems and technical systems,and their robustness is determined by all of the components inside them.Therefore,exploring the robustness of SMIS requires us to switch from a "technical + local" perspective to a "managerial + global" perspective.The modeling methods used in information systems development(e.g.,object-oriented methods,structured methods and prototyping methods)are unable to establish system models that apply to the later operation and maintenance management.Besides,trad system analysis methods(e.g.,reductionism and mechanism)are almost impossible to parse the complex behavioral characteristics presented in the operational processes of information systems.Consequently,investigating information systems needs new theories and methods.Recently,the rise of research methodologies for complex systems in the managerial domain allows us to follow research ideas with high reference value.Based on complex systems thinking,this paper synthetically utilizes synergetic theory,complex network theory,Petri net theory,etc.to study the robustness of SMIS from a managerial and global perspective.This attempt widens the application range of research methodologies for complex systems in the managerial domain,extends the research angles directing at the field of information systems,and enriches the related theoretical systems of information systems research,especially those of robustness.This paper takes different robustness-influencing factors as pointcuts to discuss the issues of robustness analysis and control.The main work and innovations are described as follows:(1)Research on the frameworks and robustness-influencing factors of SMIS.Existing literature emphasizes more on the issues in the front-end lifecycle of information systems(e.g.,system development and implementation)and less on the operation management in the back-end lifecycle.As such,there is a need to found the basic theoretical systems related to SMIS firstly.According to smart manufacturing and its demands,this paper digs deep into new or higher function and performance requirements of SMIS,analyzes the key emerging technologies as the complexity factors of SMIS.From both vertical and horizontal angles,it then dissects the composition and main links of SMIS,and builds visualized system frameworks with a certain degree of ubiquity.Finally,it discusses the robustness-influencing factors of SMIS from multiple dimensions.The discussion of those frameworks underlies the build of the following way-out models of SMIS,and the analysis of robustness-influencing factors provides chances to cognize robustness preliminarily and to select research pointcuts.(2)Robustness analysis of SMIS under interactive relationships.Previous research purely adopts engineering ways to improve the interactive relationships between information systems,and is confined to confirming the complexity of information systems.This paper focuses more on the quantitative discretion and selection of interactive relationships.Studying the mechanisms of interactive relationships on the robustness of SMIS can provide quantified references for enterprises' management practices.Specifically,it elaborates on the subsystems of SMIS and the interactive relationships between them,and constructs bidimensional nonlinear dynamic models based on adiabatic approximation theory and logistic equations.Those models take technical level and managerial level as state variables.Then,it uses stability analysis and the corresponding algorithms to disclose the operations of SMIS and the effects of interactive relationships on system robustness,and determines the threshold range of interactive relationships enabling the robust operations of SMIS.(3)Robustness analysis of SMIS under resource failures.Given that existing literature shows vagueness in studying information systems resources and emphatically studies the microscopic resource fault diagnosis,this paper anatomizes the reasonability of dividing the resources of SMIS into cyber and physical layers.Through teasing the collaborative relationships among the resources within the same layer and those among the resources within different layers,it then constructs a two-layer network model for SMIS based on complex network theory,and chooses robustness metrics that accord with the nature of SMIS.Research results demonstrate that resource failures exert significant negative effects on the robustness of SMIS.Moreover,the results find that SMIS and its cyber layer are heterogeneous systems with the "robust yet fragile" property,and assortative link is best for SMIS to resist non-random resource failures.(4)Robustness control of SMIS under data quality.Being different from previous static perception of data quality and the common-used technical strategies ameliorating data quality,this paper focuses on proposing effective control strategies to ensure the robustness of SMIS against the interference from data quality.By combining the features and demands of data from generation to application in SMIS,it explains the concrete meanings of data quality.Based on Petri net theory,it then builds process-oriented ontology for SMIS,explores the difficulty to control data nodes and task nodes and the importance of these nodes to SMIS in a quantitative manner,and devises a forward control Petri net model,a backward control Petri net model and the corresponding algorithms for system robustness,with the creation of active preventive control strategies and feedback control strategies.Lastly,it verifies the validity of the control models by simulations,and dwells on the realization of the two-stage control.The above-mentioned research is a novel attempt that applies research methodologies for complex systems to the field of management information systems,which makes the theoretical systems of SMIS research richer and more complete.Meanwhile,the research can provide scientific and down-to-earth guidance for management practices in several aspects,including improving information systems frameworks,diagnosing the health status of the operations of information systems,pinpointing internal and external interactive relationships,making effective robustness control strategies,just to name a few.