A Robust Genentic Algorithm for Automated Map Name Placement

The traditional algorithm for automated map name placement and its disadvantage are firstly analyzed in this paper.A new global optimization algorithms that called genetic algorithm is put forward to solve the point-labeling problem.According to the properties of map labeling,the basic design schemes and strategies of applying genetic algorithm to solve the map name placement is detailed in this paper.First,an integer vector coding scheme is adopted in the algorithm,which uses an array(or string) of integers to represent a configuration of map labeling.The initial strings can be obtained randomly by generating a random integer in a specified scope for every element(gene) of all strings.Then a fitness function of map labeling is represented as a quality evaluation function of map labeling Lastly some typical experiments are elaborated and some results obtained by the automated map-labeling program based on genetic algorithm are presented.In the meantime,a comprehensive experiment is conducted to compare this algorithm with climbing algorithm,annealing algorithm and Hopfield neural network method,and experimental results have shown that the performance of genetic algorithm is superior to those of other several traditional algorithms.