Lars Harrie

Cartographic generalization aims at simplifying the representation of data to suit the scale and purpose of the map. This paper deals with a method that implements the whole graphic generalization process (roughly defined as the operators simplification, smoothing, exaggeration and displacement) called simultaneous graphic generalization. This method is based on constraints, i.e. requirements that should be fulfilled in the generalization process. The constraints strive to make the map readable while preserving the characteristics of the data, which implies that all constraints cannot be completely satisfied. This study was concentrated on finding the optimal compromise between the constraints in simultaneous graphic generalization by setting weights for the constraints. Four strategies for determining the weights are described and their advantages and disadvantages are discussed. The discussion is based on the following assumptions: the constraints are independent, and the weights are only dependent on constraint type and object type. A comparison of the strategies reveals that the strategy constraint violation is the most promising. One advantage with this strategy is that it is related to the quality requirements of the map, and another advantage is that it provides a numerical measure for quality assessment. The paper concludes with a case study of the constraint violation strategy, in which visualization of the numerical quality measure is used. The case study shows that the constraint violation strategy gives a sound compromise between the constraints.