Research On Pivotal Technology Modeling Based On MAS In Maritime Aerostat Search

Along with the exploration and utilization of the ocean resource, maritime accident occurs frequently. As an important part of modern maritime safe planning, Maritime searching and rescuing is receiving more attention. As a result, it contains more tasks.The establishment of maritime searching and rescuing simulation system is for making searching and rescuing simulation training more scientific and effective, and getting the aim of enhancing the searching ability and rescuing level. On the basis of researching on the Pre-case of Tianjin Maritime Searching and Rescuing, this thesis develops simulation system with multi-Agent modeling means deeply. For better simulation effect, this simulation system brings in the concept of knowledge modeling and rules base, and explore intelligent simulation theory method, accordingly establishes knowledge model of maritime searching and rescuing system. In the process of the studying, the main points of this thesis are about the design and implement of clock Agent and weather Agent.(l) Synthesizing several kinds of time synchronizing algorithms, this thesis adopts NTP web time protocol to synchronize time. Under the consideration of the distributing character of the simulation system and the affection aroused by its network delay, it also estimates network delay by computing come and go time of the message, including: time inspecting, time getting and time synchronizing. (2) Build time Agent model, define the basic class of the Agent according to BAT model, and realize message and knowledge representation by utilizing standard communication language-FIPA ACL. (3) The realization of this weather forecasting model is done by the integration of probability producing and weather forecasting. Firstly, this thesis does statistic and analysis on weather data of Bohai Sea to get forecast factors' changing probability in different time, and then founds multiple regression models between forecast factors and predictand to get integrated forecasting. (4) Build the weather Agent model. This thesis introduces the design of weather Agent structure, the representation method of weather's factuality and consequence knowledge, and the structure flow of message harmonizing among multi Agents detailedly. Then, Lagrange interpolation technique method is brought in to solve some problems in the weather forecasting, such as availablein, valid time and anisomerous message.