This paper proposes using single-point and particle swarm operators in evolutionary programming. The single-point operator has a good global optimizing ability and accuracy, however, its convergence speed decrease. Compared to the single-point operator, the features of particle swarm operator are in the direct opposite. The two operators are complementary to each other. In this paper, the two mutation operators cooperate under a mixed strategy in order to obtain a better performance. Simulation results show that the proposed coevolutionary algorithm is superior to several popular algorithms proposed in recent years.