Research Of Distributed Framework And Task Allocation Mechanism For Marine Spatial Data Integration Based On Multi-agent

With the advent of the global informant era, the information technologyprovides a feasible technical support for deeply knowing and understanding thesea. But many kinds of approaches and the growing amount of marine data bringmarine spatial data integration the great challenge. On the one hand, the differentmarine geographical information systems causes many differences of marinespatial data for instance multiple space-times, multi-scale, multi-semantic and soon. These differences create the big problems for sharing and integrating dataamong the sea department. One the other hand, marine spatial data with complexand specialized process needs to deal with integrated automation, intelligenceand scalability. So it is the key of achieving information-based and intelligentmarine in our country how to implement the adaptation of Digital Ocean ofspatial data integration system.However at home and abroad the researches about marine spatial dataintegration technology mainly carry out against the data characteristics for eachocean area or its own application needs. Most of these studies less consider theintellectuality and expandability in the integrated processing, more emphasis onthe final results and rarely consider the efficiency. Meanwhile, in the foreignresearch we find that one of the keys of affecting the efficiency of spatial dataintegration is the issue in spatial data query option: integration tasks allocationprocess. Therefore how to quickly and efficiently find the suitable data from thedistributed of marine spatial data has become a serious problem.To solve these problems, using Multi-Agent technology, we build adistributed framework for marine spatial data integration and study themechanism for distribution of tasks. Our work mainly includes the following aspects:1) Built the framework of marine spatial data integration based onMulti-AgentThrough the analysis of marine spatial data integration process, we build theframework of marine spatial data integration by using the distributed computingand collaboration capabilities of MAS. In the framework of the system, to assigntask is the core. Following the standards, Agent queries data from the data node andsends them data processing tasks. Then these data node, which is marine departmentsthat store data, handle the marine spatial data according to local resources and theorder of these tasks. This way solve the problem about a lot of data processing andimprove the integration efficiency.2) Designed the distributed task allocation mechanism of marine spatial dataintegrationTask allocation is core in the distributed marine spatial data integration. Itincludes task analysis, decomposition, data node selection and other issues. In thestudy process, we design a distributed task allocation mechanism of marinespatial data integration by improving the traditional CNP created by Davis andSmith and add granularity of time, space, scale and other factors. Throughsimulation experiments, we prove that our mechanism has higher efficiency thanthe traditional CNP.3) Building the prototype system of marine spatial data integration based onMulti-AgentIn this paper, we build the overall structure and technical architecture of theprototype system. In the overall structure, we introduce its main modules. In thetechnical architecture, we introduce various technologies related the prototypesystem. Then, we introduce how to create all kinds of Agent classes in the JADEplatform and its communication. Finally, we display the interface of theprototype system in the application of storm surge decision support system.