进化优化与机器学习共融的水下移动无线传感器网络结构优化

水下移动无线传感器网络在海洋环境监测中具有非常重要的作用，该网络的结构优化成为学术和产业界共同关注的热点之一。但是，传感器移动和洋流导致的动态和不确定性，使得已有方法难以有效解决该问题。鉴于此，本项目研究进化优化与机器学习共融的水下移动无线传感器网络结构优化理论与方法。通过研究，拟揭示水下移动无线传感器的特性感知机理，建立水下移动无线传感器网络结构的动态不确定多目标优化模型，提出融入在线机器学习的动态不确定多目标进化优化方法，并应用于我国黄海海域生态环境监测的无线传感器网络结构优化中。该项目是自动化、人工智能、通信，以及海洋等多学科有机交叉、十分新颖且富有挑战性的研究方向，非常契合我国当前海洋强国战略的需求，产生的研究成果能够大大丰富无线传感器网络理论，大幅度提升水下移动无线传感器网络的综合性能，进一步扩大进化优化与机器学习的应用领域，因此，具有重要的理论意义和实用价值。

Underwater mobile wireless sensor networks (UMWSNs) play an important role in the ocean environment monitoring. Their structural optimization has become one of focuses in both academic and industrial communities. The dynamics and uncertainties resulted from sensor movements and ocean currents, however, make existing methods difficult to deal with the above problems. Given this, the project researches on structural optimization theories and methods of UMWSNs by combining evolutionary optimization and machine learning. It is expected to obtain the following achievements: (1) discovering the characteristic perception mechanisms of underwater mobile wireless sensors; (2) constructing dynamic uncertain multi-objective structural optimization models for UMWSNs; (3) presenting dynamic uncertain multi-objective evolutionary optimization methods integrating online machine learning; (4) applying to the structural optimization of UMWSNs in the ecological environment monitoring of the Yellow Sea. This project, at the cross edge of such disciplines as automation, artificial intelligence, communication and ocean, is very novel and challenging, as well as meets the demand of maritime power strategy. The achievements of this project can enrich the theory of wireless sensor networks, greatly enhance the comprehensive performance of UMWSNs, and enlarge application areas of evolutionary optimization and machine learning; thus it is considerable significance in theory and application.