optimal control via neural networks : a convex approach
读后感:
我们都知道深度神经网络,一般来说输出对于输入都是非凸的,包含很多局部最优点,所以在优化过程中很容易陷入局部最优情况。在对稳定性要求很高的系统控制情境下(比如电力系统控制,航天系统以及工业控制),这种多个局部最优解并且没有全局最优收敛性保证的情况是我们非常不愿看到的,也一定程度限制了目前深度模型在这些行业中的应用。
yet few research investigated how to integrate deep learning models into real-time closed-loop control of physical systems.
Neural networks, because of their structures, are generally not
convex from input to output. Therefore, many control applications (e.g., where real-time decisions
need to be made) choose to favor the computational tractability offered by linear models despite their poor fitting performances.
该算法不仅有良好的预测准确率还有良好的计算优化性能。
我们的工作的另一个主要贡献是,我们明确证明ICNN可以表示所有的凸函数和系统动力学,并且比广泛使用的凸分段线性逼近具有指数更高的效率。
The network shown in Fig. 2(b) is a convex function from inputs t