Research And Realization Of Disease Propagation Oriented Artificial Society Meta-modeling

The outbreak of epidemic has become more and more frequently in the last few years, and as a result of the development of the society, its propagation trends to be more complex、networked and socialized. When dealing with the intervention of the epidemic, the public health department faces tremendous difficulties, which urgently calls a dynamic simulation system based on the ―scenario-response‖ theory which can be customized、computed and extended. As a method extensively using in complex systems, artificial society has become a main approach to study the complexity of the human society, meanwhile, it is promising in the study of epidemic propagation.Artificial society is an inter discipline subject, which means its modeling work must take the character of multi-fields into account. New modeling methods that can support the requirement of the modeling work and its inter operate among models are urgently needed. Domain Specific Modeling(DSM) based on Meta-modeling and Model Transformation offers an available and nice choice for the problem.Under the background of the project ―Parallel Emergency Management based Integrated Distillated Platform for Dynamic Simulation and Computation Experiments of Emergency Events‖, this article suggested a disease propagation oriented artificial society meta-modeling approach, and realized a public health emergency case study, combined with the generic modeling environment(GME). The main works of this article are:1) Based on the method of meta-modeling, suggested the disease propagation oriented artificial society domain modeling language. This article deeply discussed the basic concept of Model Driven Architecture(MDA) 、 Meta-Modeling 、 Domain Modeling Language(DML) and Model Transformation. Based on the analysis of the character and bottle-neck of the modeling of artificial society, this article designed the meta-modeling based artificial society domain specific modeling language. The language chooses the meta-models and extend object constraint language(OCL) to describe the abstract syntax, meanwhile, emploees graphical icons to model the concrete syntax.2) Realized the sematic modeling by model transformation, suggested three types of model transformation methods. This article designed and realized the model transformation both from artificial society concept models to formalism modeling diagrams and from finite state machine to codes. Considering the practices of the engineering, the third type of model transformation was suggested: from domain specific models to codes directly. These three model transformation types can describe the sematic modeling effectively, and ensure the high efficiency、maintenance、coherence of the system codes.3) Designed several familiar intervention models for the disease propagation, and have done some computational experiments. Aiming at the essential goal which is to support the decision making while confused by the intervention of epidemic, this article integrated several intervention models into the disease propagation oriented artificial society, such as vaccination、isolation and public space closing. Comparing with the traditional approaches, the simulation approach can quantificationally put out more detailed results, which allows people to estimate the effect of the interventions quantificationally.On one hand, this article successfully applied the meta-modeling method to the modeling procedure of the disease propagation oriented artificial, which is the base work of the simulation approach to test the effect of the interventions, meanwhile, it enriches the basic modeling methods of the simulation subject. On the other hand, the meta-modeling based domain specific modeling language promote the extensibility、reusability、interoperability and the capability of the integration of the models.