An improved adaptive genetic algorithm is presented in this paper. It primarily includes two modified methods: one is novel adaptive probabilities of crossover and mutation, the other is truncated selection approach. This algorithm has been validated to be superior to the simple genetic algorithm (SGA) by a complicated binary testing function. Then the proposed algorithm is applied to optimizing the planar retrodirective array to reduce the cost of the hardware. The fitness function is discussed in the optimization example. After optimization, the sparse planar retrodirective antenna array keeps excellent retrodirectivity, while the array architecture has been simplified by 34%. The optimized antenna array can replace uniform full array effectively. Results show that this work will gain more engineering benefits in practice. An improved adaptive genetic algorithm is presented in this paper. It was included two modified methods: one is novel adaptive probabilities of crossover and mutation, the other is truncated selection approach. This algorithm has been validated to be superior to the simple genetic algorithm (SGA ) by a complicated binary testing function. Then the proposed algorithm is applied to optimize the planar retrodirective array to reduce the cost of the hardware. The fitness function is discussed in the optimization example. After optimization, the sparse planar retrodirective antenna array keeps excellent retrodirectivity , while the array architecture has been simplified by 34%. The optimized antenna array can replace uniform full array effectively. Results show that this work will gain more engineering benefits in practice.