Abstract:The paper introduces a mathem atical model of testpaper assembling on genetic algorithm, defines an adaptive function on testpaper assembling, and provides some ideas on multi-object parameter optimization on restricted terms by genetic algorithm. In the evolutionary processes of seeds initialization, operators selecting, operator crossing, operation differentiation, the best solution is finally worked out. Results of experiments indicate, genetic algorithm is more effi cient than other algorithms on testpaper auto-assembling.Random algorithm which could achieves testpaper non-adjacency distribution, is a new method for similar multi-object restriction and non-adjacency combination problems.