论文部分内容阅读
反应动力学模型往往是非线性多极值函数,对其参数优化,传统的确定性优化算法易陷入局部最优。全息搜索策略(HRS)是一种寻优效率较高的确定性优化算法,但只能用于离散系统的优化。通过对连续变量进行离散化处理,并运用迭代计算逐步缩小离散系统与原连续系统的偏差,将复杂的多维连续变量优化问题转化为多个串联的较为简单的离散变量组合优化问题,使HRS只需在有限个离散解中寻优,实现HRS的连续变量优化,并在HRS确定性寻优过程中引入随机性交叉操作,以进一步提高算法的全局搜优效率,由此建立了一种改进的HRS。八维Alpine函数测试表明,改进的HRS的全局优化性能明显优于常规遗传算法(SGA),将其应用于2,3-二氟-6-硝基苯酚的合成反应动力学参数优化,得到的动力学方程与原方程相比,拟合偏差平方和减小了17.2%,验证的平均相对误差由5.37%降至3.44%,参数优化结果较为满意。
The reaction kinetic model is often a nonlinear multipole value function. For its parameters optimization, the traditional deterministic optimization algorithm tends to fall into the local optimum. Holographic search strategy (HRS) is a deterministic optimization algorithm with high efficiency, but it can only be used for optimization of discrete systems. By discretizing the continuous variables and using the iterative computation to gradually reduce the discrepancy between the discrete system and the original continuous system, the complex multi-dimensional continuous variable optimization problem is transformed into a series of relatively simple discrete variable combinatorial optimization problems that make the HRS only It is necessary to optimize in a finite number of discrete solutions to achieve continuous variable optimization of HRS and to introduce a randomized crossover operation in the deterministic optimization of HRS to further improve the overall search efficiency of the algorithm and thus to establish an improved HRS. The eight-dimensional Alpine function test shows that the global optimization performance of the improved HRS is obviously superior to that of the conventional genetic algorithm (SGA), and the optimized kinetic parameters of 2,3-difluoro-6-nitrophenol are optimized. Compared with the original equation, the kinetic equation decreases by 17.2% and the average relative error of validation decreases from 5.37% to 3.44%, and the parameter optimization results are satisfactory.