Connected cyber and physical elements exchange information through feedback in a cyber-physical system (CPS). Since CPS oversees the infrastructure, it is an integral part of modern living and is viewed as crucial to the development of cutting-edge smart devices. As the number of CPSs rises, so does the need for intrusion detection systems (IDS). The use of metaheuristic methods and Artificial Intelligence for feature selection and classification can offer solutions to some of the problems caused by the curse of dimensionality. In this research, we present a blockchain-based approach to data security in which blocks are generated using the RSA hashing method. Using Differential Evolution (DE), we first select the blockchain-secured data, and then we partition that data into train and testing datasets to use for training and testing our model. It is also permitted for the validated model to use a deep belief network (DBN) to predict attacks. The purpose of the simulation is to evaluate the safety and precision of the classifications. It turns out that the proposed strategy not only improves classification accuracy but also makes the data more resistant to attacks.

连接的网络和物理元素通过网络物理系统（CPS）中的反馈交换信息。由于 CPS 负责监督基础设施，因此它是现代生活不可或缺的一部分，并且被视为对尖端智能设备的开发至关重要。随着 CPS 数量的增加，对入侵检测系统 (IDS) 的需求也在增加。使用元启发式方法和人工智能进行特征选择和分类可以为一些由维数灾难引起的问题提供解决方案。在这项研究中，我们提出了一种基于区块链的数据安全方法，其中使用 RSA 哈希方法生成块。使用差分进化（DE），我们首先选择区块链保护的数据，然后将该数据划分为训练和测试数据集，以用于训练和测试我们的模型。经验证的模型还允许使用深度信念网络（DBN）来预测攻击。模拟的目的是评估分类的安全性和精度。事实证明，所提出的策略不仅提高了分类精度，而且使数据更能抵抗攻击。