Research on Model Beginning Standard of Multi-objective Evolutionary Algorithm Based on Entropy
Qiong Yuan, Guangming Da · Journal of Software · 2016
Entropy is a measurement of the system's degree of chaos, it has important applications in the field of control theory, probability theory, number theory, astrophysics and life sciences.In the process for multi-objective evolutionary algorithm to solve optimization problem, the distribution of individuals in the population goes from disorder to order.The paper introduces the entropy theory to multi-objective evolutionary and proposes an entropy calculation method to describe the population distribution, besides, it offers a method to judge the critical point, which can be used as a model beginning standard.