复杂动力学模型参数优化问题的协同进化算法

Cooperative Co-Evolutionary Algorithm for Optimization of Kinetic Parameters

复杂反应动力学建模中,系统参数的优化是需要解决的关键问题之一。该类优化问题具有多参数、非线性以及参数相关性强等特点。协同进化算法将多种群之间的协同作用以及种群内部的独立进化相结合,适合于求解该类问题。将改进的协同进化算法应用到化工氧化反应建模过程的系统参数优化问题中,避免了解决该类问题的传统优化算法中易陷入局部极值以及初值依赖性强的缺点,运用理论证明了该算法的有效性。测试结果表明,协同进化算法对于求解该类复杂参数优化问题是有效的。

Parameters optimization is one of the key problems in complex reaction kinetic modeling.This problem has many nonlinear parameters with high relevance.Combined the cooperative function among subpopulations and the independent evolution in subpopulation,the cooperative co-evolutionary algorithm is fit for this kind of problems.Applying the cooperative co-evolutionary algorithm into the problem of parameters optimization in oxidation reaction modeling,the efficient of algorithm is proved theoretically.Experimental results also demonstrate that the algorithm is fairly efficient for solving the optimization problem of parameters.