最陡下降-模拟退火算法求解镉-腐殖酸配合物累积稳定常数

Calculation of the Cumulative Stabilization Constants of Cd-HA Complexes by Gradient Descent-simulated Annealing Algorithm

采用阳极溶出伏安法结合人工神经网络法中的最陡下降-模拟退火算法，研究了在碱性介质中镉与天然水环境中的有机物质腐殖酸（HA）的配合物，用ASV法测定了来源于煤炭腐殖酸的富里酸分子量平均值为995．2，求得PH≌10条件下的Cd2+-HA的1级和2级配合物累积稳定常数值1gβ1=4．079，1gβ2=8．602．

In this paper, the cumulative stabilization constants 0f Cd-HA (Humic Acid) inthe alkali media have been studied and calculated using anode stripping voltammetry method(ASV) cornbined with GD-SA algorithm in artificial neural network(ANN)- The molecularweight of humic acid originated from coal was determined by ASV method and its averagevalue is 995. 2, The cumulative stabilization constants of the first and second order complexes of Cd2+-HA under the condition of pH = 10 were calculated and their values are 1gβ1 =4. 079 and 1gβ2=8. 602 respectively. Two local minimum of objective function appear in iterative process. The values of objective function jump out from these local minimum successfully and reach to the global optimization.