(1)针对在线计算量大这一缺陷，将预测控制中的柔化输出信号的思想推广到柔化输入信号，使得约束条件被简化为仅对当前控制量的约束，可以直接计算得出；同时该方法避免了求逆矩阵，大大减小了计算量，并能够保证控制算法的可行性和良好的控制性能。

(2)针对传统算法中设计参数整定困难这一缺点，应用基于BP神经网络变参数设计的广义预测控制算法，实现了对控制量柔化参数的在线调整。

(3)利用带有遗忘因子的最小二乘法对系统辨识。本文通过仿真发现该方法对于Hénon混沌系统并不完全适用，可考虑利用其他优化系统辨识的方法对本方法进行改进，以期达到更好的辨识效果。

(4)针对系统稳定性分析复杂，本文在控制增量前加入前馈因子，保证所选的Lyapunov函数使闭环系统满足Lyapunov稳定判据，由此证明闭环系统稳定。

-1. To solve the problem of GPC huge computation, algorithm with input increment constraints is presented in which the concept of output softness was used to soften the input increments．As a result, the constraints are simplified to be the only one constraint on the current control increment which can be computed directly. At the same time, it needn’t computing the inverse matrix and thus reduces large computation. Moreover, it guarantees the feasibility of the algorithm and has good control performance．

2. To overcome the difficulty in the choice of tuning parameters in traditional GPC, a GPC algorithm with variable parameter design based on BP neural network. is presented，in which the input softness parameters are tuned on line.

3. In this paper, we Identify system by using the least square method with forgetting factor. However, after system simulation, we realize that this method doesn’t fit the Hénon chaotic system perfectly. So we recommend modify this method by other Optimizati