A DYNAMIC DECISION TREE STRUCTURE SUPPORTING URBAN ROAD NETWORK AUTOMATED GENERALIZATION

Generalization is a complex task which requires a good understanding o f the geometrical and semantic aspects of map features as well as the potential use of the map. Reasoning about spatial relationship, compar ing alternative solutions, and contextual thinking are all important a ctivities required for generalization decision-making. Such activities are not easy to simulate in a computer without sufficient information and the support of good data structures and adequate reasoning mechan isms. This paper introduces a dynamic decision tree structure in an at tempt to partly circumvent the problem of urban road network generaliz ation through the use of object classification and aggregation hierarc hies, topological data structure, decision rules, and AI technology. A part from the construction process, it also discusses the reasoning pr ocess for decision-making, and provides two test results with some dis cussion on the benefits and short-comings of the proposed approach.