An Analysis of the Micro-Macroscopic Interaction in Evolutionary Computation and its Application to Economical Systems

本研究的目的是建立一种涉及个体间相互作用的进化计算新模型。由于在传统的电子商务模型中，每个个体都是独立评估的，因此它们在获取人工市场策略时并不有效。本文介绍了邻域评价方法。这种方法不仅评估健身景观中感兴趣点的个体，还评估邻近点的个体。实证结果表明，邻域评价优于传统评价。

The objective of this work is to develop a new model of evolutionary computation (EC) which involves a kind of interaction among individuals. Since each individual is evaluated independently in conventional EC models, they are not effective in acquiring strategies for artificial markets. In this work, the neighborhood evaluation is introduced. This method evaluates not only the individual at the point of interest in the fitness landscape but also the neighboring ones. The empirical results show that the neighborhood evaluation outperforms the conventional one.