论文部分内容阅读
针对最小二乘支持向量机(LSSVM)在建模中的重要参数如何选择问题。提出利用具有随机性、遍历性及规律性的混沌优化算法对LSSVM建模过程中的参数进行优化搜索,为了加快对较大搜索空间中的搜索速度,提出变尺度混沌优化算法与遗传算法相结合遗传算法的组合算法对LSSVM中的参数优化。组合算法克服了单一算法存在的早熟、局部收敛及寻优速度慢等问题,把混沌变量种群映射到LSSVM参数取值区间,按照遗传算法训练,同时利用训练集训练LSSVM,最终得到参数优化值。将该方法应用的谷氨酸发酵过程的建模研究,取得了较高建模精度,提高发酵过程资源利用率的同时增加了谷氨酸产量。
How to select the important parameters in the modeling of Least Squares Support Vector Machine (LSSVM). In order to accelerate the search speed in large search space, a chaos optimization algorithm with randomness, ergodicity and regularity is proposed to optimize the parameters in LSSVM modeling process. A chaos optimization algorithm based on scaling is combined with genetic algorithm Optimization of Parameters in LSSVM by Combination Algorithm of Genetic Algorithm. The combinatorial algorithm overcomes the problems of premature, local convergence and slow optimization of the single algorithm. The chaotic variable population is mapped to the LSSVM parameter interval and trained according to genetic algorithm. At the same time, the training set is used to train LSSVM to get the optimal parameter finally. The modeling of the glutamic acid fermentation process applied by the method has achieved higher modeling accuracy, increased resource utilization rate in the fermentation process and increased glutamic acid production.