Basic Research on the Transformation of Market Structure

首先，收集了有关产业结构变化的数据。尽管我不顾日本国内外的各种相关团体，查阅了尽可能多的资料，但遗憾的是，我不禁判断，没有足够的数据(月度数据)来以具有统计意义的方式分析市场结构的变化，由于数据的收集工作花费了相当多的时间和精力，我不得不把理论模型构建工作放在理论模型构建上，只专注于模拟模型的改进。到目前为止，我已经介绍了(1)使用转移概率来描述消费者行为的统计物理技术和(2)描述企业行为的强化学习算法。以下是未来的工作：a)内生化一个关于消费者更换企业购买商品的转移概率的重要参数，以便受到企业广告的影响；b)引入或改进代理人的学习过程，使他们的行为更具智力(但以有限理性的方式)。c)通过广告投资等方式纳入品牌战略。

First of all, data concerning the change in the industrial structure was collected. Though I inquired of various related groups regardless of the inside and outside Japan and examined material as far as I could do, I could not help judging, to my regret, that enough data (monthly data) does not exist to analyze changing the market structure in the statistically meaningful manner.Since the considerable time and the effort were used for the collection work of data, I had to put over theoretical model construction and concentrated only on the improvement of the simulation model. Up to now, I have introduced (1) a statistical physics technique that uses the transition probability to describe consumer behavior and (2) an algorithm of reinforcement learning to describe firm behavior.The following are regarded as future works :a)An important parameter concerning consumer's transition probability of changing firm to purchase goods is endogenized so as to be affected by the advertisement of firms.b)Learning processes of agents are introduced or refined so that they behave more intellectually (but in the boundedly rational manner).c)The branding strategy through an advertising investment etc. is incorporated.