Intelligent test sheet generation is a multi-objective constrained optimization problem. Genetic algorithm based on groups search strategy can provide a better solution for multi-objective optimization. Traditional genetic algorithm in test sheet generation process has many drawbacks, such as poor convergence, low fitness and high exposure times. To solve these problems, this paper proposes an adaptive cellular genetic algorithm based on selection strategy. Selection strategy can adaptively determine candidate test items set and the conceptual granularities according to the desired concept scope. Then, a new cellular population is formed by candidate test items. After evolution by the rule, genetic algorithms are executed. The experimental results show that the proposed algorithm gets rid of tests that do not meet the requirements which can reduce knowledge related errors, lower the exposure of tests, and increase the possibility of escape from local optima. In general, the algorithm proposed in this paper effectively improves the convergence speed as well as generates test papers more in line with people's demands.

No abstract