역사 시뮬레이션 구축의 제 문제와 그 방법에 대한 연구

Abstract

The aim of this study is examine of various difficulties of computational history. Computational history is a history that uses calculation as its main research method. Here, calculation refers to simulation. There are a lot of difficulties of introduction of computational history in historical research. This study examine three problems. That are simulation and historical coincidence, simulation and historical representation and method of the simulation. First, history handles coincidence. simulation can either. Coincidence is a key concept in simulation. This can be seen from Buffon's thought experiment. The only difference simulation has is that simulation coincidence is subject to purpose. This is due to the fact that simulation is basically an algorithm for solving problems. Historians can use this nature of simulation to approach historical problems. Second, the issue of representation. History is a representation of the facts of the past, and simulation is an imaginary representation. However, in essence, representation is a representation of an idea. This can be seen by looking at the example in the photograph. Photograph that we consider to be representations of reality are actually only mechanical representations of ideas(moment). However, we derive the usefulness of reality from the representation of ideas. Likewise, we explore historical facts from historical representations, that is, representations of ideas. This is not much different from simulation, which is a representation of an idea. Therefore, we do not need to have any doubts in using simulation in our historical exploration. Finally, there are two main methods of simulation. System dynamics and agnet-based simulation modeling(ABM). In this study I proposed an agent-based simulation modeling as a method for historical simulation. ABM is a computational model that seeks to understand the entire system based on autonomous agents. At this time, it is assumed that the agents communicate with other agents and act in order to achieve a specific goal within limited autonomy with their own attributes and their own rationality. This ABM has the advantage of not only being able to build a model that encompasses the system and its members, but also analyzes both the interrelationships between members and the interrelationships between members and systems. Simulation platforms to build ABM include Refast, Mason, Netlogo, and Mesa. In particular, Mesa has the advantage of being able to design a large number of agents and considering spatial factors. And there is a trend of gradually appearing historical researches using the ABM platform. Based on this research, I expect more studies on the history of simulation to emerge.