农作物水量的智能分配与预测研究

发布时间：2018-07-29 17:25

【摘要】：在现代农业的发展过程中,水资源分配不均,供需矛盾的问题日益加剧,如何发展节水型农业是亟待解决的一个重要问题。除了采用成熟的微灌、喷灌、滴灌等相对先进的灌溉措施之外,在实际的农业生产过程中,应该实施合理有效的灌溉制度,以农作物实际需水量为依据,通过对大量相关数据的分析,应用先进的技术实施精确灌溉,以此来提高灌溉的效率,提高水的利用率。如何在现有水资源的基础上,合理有效地进行农作物的灌溉具有十分重要的现实意义。智能灌溉以大量的数据分析为基础,改变了以往水资源灌溉过程中的盲目性和随意性,能够使管理成本不断降低,使得经济效益明显提高。本论文结合农业物联网的实际背景,以不同农作物水资源的优化调度和预测为基础,主要开展了以下工作:(1)针对传统免疫优化算法存在的不足,通过加入局部搜索算子,改进了传统的免疫优化算法;同时,为了加快种群迭代的速度,防止算法在迭代过程中可能会错过最优抗体解的情况,将初始种群分成了两个子种群进行并行搜索,在一定程度上加快了种群搜索的速度。(2)基于实际农田中的玉米和小麦不同生长周期的情况,在水资源供给充足的情况下,验证了上述改进的免疫优化算法相比于原免疫优化算法的优势。同时在非充分灌溉的条件下,通过改进的免疫优化算法,协调了两种农作物不同生长周期的水量分配,使得两种农作物的总产量达到最大。(3)针对农田灌溉过程中水资源分配不合理和浪费的现状,结合上海市农委官网提供的数据,以最小二乘支持向量机为基础,对于传统的以经验数据来选择支持向量机的两个重要的参数C和?,可能存在预测不够准确的情况,本文将改进后的粒子群算法和免疫优化算法分别应用到最小二乘支持向量机的参数优化中,形成新的支持向量机模型。(4)以实际采集到的农业数据为基础,将上述的支持向量机模型运用在实际农作物的水量预测中,通过影响农作物需水量的各种因素之间的相互作用,达到预测农作物在某一时刻的需水量的目的。通过比较两种智能算法,验证了本文提出的改进的免疫优化算法和改进后的粒子群算法在参数优化方面的作用,同时证明了提出的改进的免疫优化算法比粒子群算法的优势,这些可以为后续农作物的节水灌溉提供一定的理论指导。最后总结了全文的内容并对未来可以研究的内容进行了展望。
[Abstract]:In the course of the development of modern agriculture, the distribution of water resources is uneven and the contradiction between supply and demand is becoming more and more serious. How to develop water-saving agriculture is an important problem to be solved urgently. In addition to adopting relatively advanced irrigation measures such as mature micro-irrigation, sprinkler irrigation and drip irrigation, a reasonable and effective irrigation system should be implemented in the actual agricultural production process, based on the actual water demand of crops. Through the analysis of a large number of related data, the advanced technology is applied to carry out accurate irrigation, so as to improve the efficiency of irrigation and the utilization rate of water. How to reasonably and effectively irrigate crops on the basis of existing water resources is of great practical significance. Based on a large amount of data analysis, intelligent irrigation has changed the blindness and arbitrariness in the process of irrigation of water resources in the past, which can reduce the management cost and increase the economic benefit. Based on the practical background of the agricultural Internet of things and the optimal scheduling and forecasting of different crop water resources, this paper mainly carried out the following work: (1) aiming at the shortcomings of the traditional immune optimization algorithm, by adding the local search operator, The traditional immune optimization algorithm is improved, and in order to speed up the population iteration and prevent the algorithm from missing the optimal antibody solution in the iterative process, the initial population is divided into two sub-populations for parallel search. (2) based on the different growth cycles of corn and wheat in real farmland, when the water supply is sufficient, The advantages of the improved immune optimization algorithm compared with the original immune optimization algorithm are verified. At the same time, under the condition of inadequate irrigation, through the improved immune optimization algorithm, the water allocation of two crops with different growth cycles is coordinated. The total yield of the two crops is maximized. (3) in view of the unreasonable and wasteful distribution of water resources in the process of irrigation, combining with the data provided by the official website of Shanghai Agricultural Commission, the least square support vector machine is used as the basis. For the traditional selection of two important parameters C and C of support vector machine based on empirical data, the prediction may not be accurate enough. In this paper, the improved particle swarm optimization algorithm and the immune optimization algorithm are applied to the parameter optimization of the least squares support vector machine, and a new support vector machine model is formed. (4) based on the agricultural data collected, The support vector machine (SVM) model is used to predict the water demand of crops at a certain time through the interaction of various factors affecting crop water demand. By comparing two kinds of intelligent algorithms, the function of the improved immune optimization algorithm and the improved particle swarm optimization algorithm in parameter optimization is verified. At the same time, the advantage of the improved immune optimization algorithm is proved to be better than that of the particle swarm optimization algorithm. These can provide certain theoretical guidance for the subsequent crop water-saving irrigation. Finally, the content of the paper is summarized and the future research content is prospected.
【学位授予单位】：东华大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：S27;TP18

【相似文献】