Abstract: For reasonably and justly selecting evaluation experts from the alternative expert database, the information entropy is introduced as a measure degree of order in the system according to expert selection rules. According to rules, computational constraint conditions are set up based on information entropy, and the probability distribution of expert selection is computed by Lagrange operators. From the practical alternative expert database, the real data of selected experts is extracted to execute computational experiments, the result of annual experts selection shows the random feature, which proves that the method can avoid the cycle-correlated issue of conventional random generators and manage to select experts evenly.

Key words: non-equilibrium, maximum entropy, entropy increasing, constraints, Lagrange operator

摘要： 为合理、公平地从备选库均匀抽取评审专家，根据专家挑选规则引入信息熵测度作为体系中专家挑选的有序度量，依据规则建立基于信息熵的计算约束条件，利用极大熵理论构造专家挑选概率分布函数，并利用拉格朗日算子计算专家挑选的全概率分布，从实际备选库抽取真实的专家挑选数据进行计算实验，年度专家挑选结果呈现随机性质，证明该方法能克服常规随机发生器的周期自相关性，实现专家的均衡挑选。

关键词: 非均衡, 极大熵, 熵增加, 约束条件, 拉格朗日