Wang WenBin, Wu ZiNiu, Wang ChunFeng, Hu RuiFeng
1Department of Engineering Mechanics, Tsinghua University, Beijing, 100084 China.
2School of Electro-Mechanical Engineering, Xidian University, Beijing, 100084 China.
Sci China Phys Mech Astron. 2013;56(11):2143-2150. doi: 10.1007/s11433-013-5321-0. Epub 2013 Oct 3.
A model based on a thermodynamic approach is proposed for predicting the dynamics of communicable epidemics assumed to be governed by controlling efforts of multiple scales so that an entropy is associated with the system. All the epidemic details are factored into a single and time-dependent coefficient, the functional form of this coefficient is found through four constraints, including notably the existence of an inflexion point and a maximum. The model is solved to give a log-normal distribution for the spread rate, for which a Shannon entropy can be defined. The only parameter, that characterizes the width of the distribution function, is uniquely determined through maximizing the rate of entropy production. This entropy-based thermodynamic (EBT) model predicts the number of hospitalized cases with a reasonable accuracy for SARS in the year 2003. This EBT model can be of use for potential epidemics such as avian influenza and H7N9 in China.
提出了一种基于热力学方法的模型，用于预测假定受多尺度控制努力支配的传染病动态，从而使系统与一个熵相关联。所有的疫情细节都被纳入一个单一的、随时间变化的系数中，该系数的函数形式通过四个约束条件确定，其中特别包括存在一个拐点和一个最大值。求解该模型得到传播率的对数正态分布，为此可以定义一个香农熵。唯一表征分布函数宽度的参数通过最大化熵产生率唯一确定。这个基于熵的热力学（EBT）模型对2003年非典住院病例数的预测具有合理的准确性。该EBT模型可用于中国潜在的疫情，如禽流感和H7N9。
