An abstract mathematical framework for semantic modeling and simulation of urban crime patterns

摘要

Crime is not random. Criminologists contend there is predictable rationality and definite patterning behind urban crime. Conventional research for crime analysis is statistical and empirical in nature. However, with increasing complexity of the involved sociological system, empirical deduction is not sufficient; mathematical and computational models are needed for reasoning about system dynamics. In this thesis, we posit a novel approach of computational modeling of urban crime patterns. By combining the Abstract State Machine (ASM) formalism with the Multi Agent System (MAS) modeling paradigm, we obtain an abstract formal framework for semantic modeling and integration of established theories of crime analysis. Such a firm mathematical foundation also provides a quintessential platform for constructing discrete event simulation models. The framework can be applied for predictive and explanatory modeling of crime patterns. The virtue of this work is in its pioneering nature. It introduces an unprecedented, interdisciplinary research field of Computational Criminology.