Collaborative Map Generalization Method Of Contours And Rivers Based On Multi-Agent

Since the1960s, map generalization has always been focused on in the academic and industrial domains. Map automatic generalization in digital environment is still one of the international problems and also hot topic in modern cartographic research. With the rapid development of the computer and artificial intelligence technology, all geographical feature types, whole process and completely control have become the ultimate goals of map automatic generalization, which promoted the development of the multi-features collaborative generalization. Relative researches about the landform, river system, road, settlement appeared at domestic and abroad. Due to the close relationship forming from the geological evolution, landform and river system became the preference in map generalization research.Through analyzing the spatial relationships between contours and rive features in advance, the strategy of collaborative generalization on contours and rivers had been proposed, which reduced the spatial conflicts between contours and rivers after generalization, however, this approach was not flexible and lacked of decision mechanism, it needed to design specific solutions before the generalization process, once the scheme was determined, the unique generalization result could be found. The emergence of Agent technology provided new ideas to collaborative generalization of contours and rivers. Agent technology had abilities of autonomy, reaction and sense of environmental changes. Supported by communication between the multi-agents, complex and dynamic problems could be solved. Using multi-agents to solve map generalization task could better simulate human’s generalization process. Through integrating cooperative rules into the multi-agent collaborative generalization, the collaborative scheme could be more flexible. Diverse results could be gained through the different choice of knowledge rules, so as to achieve real intelligent of the generalization process.This paper was intended to implement the collaborative generalization of the contour lines and rivers utilizing the Agent technology. Main points of this research are as follows:First, based on investigation and study of the generalization norms, principles and methods of contours and rivers, the map generalization knowledge classification system was established. Three types of map generalization knowledge were proposed:the description knowledge, the judgment knowledge and the procedural knowledge, and thus formal expression and usage methods have been discussed; Second, specific rules and collaborative mechanisms in the process of generalization about contours and rivers were researched. Cooperative relationships between contours and rivers were discussed and the relationship extraction method was analyzed, an automatic watershed extraction approach with collaboration of contours and rivers was proposed; In order to take advantage of agent technology on the collaborative generalization of contours and rivers, certain types of domain agents were defined, and the agent model with specific inner architecture was designed and the inner operating mechanism was also described in detail. Third, the river selection knowledge and the cooperative simplification rules of contours and rivers have been summarized; river selection reasoning process and cooperative simplification process have been analyzed. DLG data of contours and rivers in large scale were taken as an experiment data source. An experimental platform was built on the basis of Eclipse3.5and JADE. The collaborative generalization experiment in contours and rivers was performed. Experimental result showed that collaborative generalization method based on multi-agent was flexible and intelligent, the prospective purpose and requirements were achieved.