An algorithm of constrained generalized predictive control was proposed based on partical swarm optimization(PSO) and simulated annealing(SA).PSO and SA were used for iterrative optimization to enhance the search capabilities for constraint boundary. Meanwhile, a violation function was constructed by constraints. This function and fitness function  determined the optimal solution better or not. Constraint handling performance of GPC can be effectively improved by this algorithm. Its feasibility and efectiveness are verified via simulation study for an industrial object．

Generalized predictive control,   ,  Constraint,   ,  Particle swarm optimization,   ,  Simulated annealing