Unconventional emergencies usually have significant influence on safety andstability of human societies. As a result, it is particularly important to effectively predictand recognize unconventional emergencies and their evolution regularities, especially indecision-making and emergency management. Conventional predict-handle approach toemergency management is deficient in dealing with the challenges of modeling,analysing, and managing unconventional emergencies, not to mention meeting therequirements of drilling and plan-evaluation oriented to unconventional emergencyscenarios. ACP (Artificial society, Computational experiments, and Parallel control)methodology facilitates the administration of emergency scenario drilling and scientificdecision-making by constructing and experimenting on an artificial society which isparallel to the real society. Artificial society provides a means to model and simulate thereal society, and serves as a basis of ACP methodology.Artificial society modeling requires specification of individuals, environments, andsocial relations in a society, which involves great intricacy. In order to dynamicallysimulate emergencies, an artificial society model should support emergencyspecification and simulation. Such a model is usually platform-independent,necessitating the generation of simulation instances. Rapidly generating models andinstances is therefore critical for improving efficiency and effectiveness ofcomputational experiments and parallel control.This paper proposes ASML, a language for modeling artificial society, anddiscusses the relationship and transformation rules between artificial society models andsimulation instances. Algorithms implementing these rules are presented. A toolsupporting ASML-based modeling is implemented. A case study is presented in order todemonstrate the validity of the tool. |