论文部分内容阅读
在水产养殖业中,水质对水中生物体的生长具有重要的影响,影响水质的因素主要包括:养殖水体的温度、pH值、氨氮含量、水中的溶解氧含量等等。根据统计资料显示,直接或者间接的遭受缺氧致死的鱼类,大约占到养殖鱼类死亡总数的60%,因此对水质溶氧含量进行预测对水产养殖业具有很大的意义。在预测方面,传统神经网络容易陷入局部最优,模型的推广能力不够强,支持向量机模型能够克服神经网络的这个缺点,具有很好的推广能力。本文运用变尺度混沌量子粒子群优化算法优化最小二乘支持向量机,选取国家罗非鱼产业技术研发中心无锡养殖基地的实际测量数据作为训练和测试样本数据,对水质溶氧情况进行预测。针对粒子群优化算法和量子粒子群优化算法容易陷入早熟的缺点,提出变尺度混沌量子粒子群优化算法来对最小二乘支持向量机进行参数寻优,并将这种建模方法运用于水质溶氧预测中。将传统神经网络模型以及基于量子粒子群优化算法优化的最小二乘支持向量机模型的预测结果与本文所建立的模型的预测结果相比较,证明了本文算法具有优越性,同时该模型较好的预测了水质溶氧趋势,为渔业的养殖提供了良好的参考价值。
In the aquaculture industry, water quality has an important impact on the growth of aquatic organisms. Factors that affect water quality include: temperature, pH value, ammonia nitrogen content, dissolved oxygen content in water and so on. According to statistics, fish that are directly or indirectly killed by hypoxia account for about 60% of the total deaths of farmed fish. Therefore, forecasting dissolved oxygen content in water quality is of great significance to aquaculture. In the prediction, the traditional neural network is apt to fall into the local optimum, the promotion ability of the model is not strong enough, and the support vector machine model can overcome the shortcoming of the neural network and has good promotion ability. In this paper, a variable-scale chaotic quantum-particle swarm optimization algorithm is used to optimize the least-squares support vector machine. The actual measurement data of the national tilapia industry R & D center Wuxi breeding base is selected as training and testing sample data to predict the dissolved oxygen in water. Aiming at the shortcomings of PSO and QPSO, it is proposed to use variable-size chaotic quantum particle swarm optimization algorithm to optimize the parameters of LS-SVM, and to apply this modeling method in water quality Oxygen prediction. Comparing the prediction results of the traditional neural network model and the least-squares support vector machine model based on the optimization of quantum particle swarm optimization algorithm with the prediction results of the model established in this paper, it is proved that the proposed algorithm is superior and the model is better The trend of dissolved oxygen in water quality is predicted, which provides a good reference value for fishery breeding.