Proposed Security Evaluation of Post-Quantum Cryptography Based on Soft Computing

Developing security evaluation algorithms is one of computer networks' most significant difficulties today. This is a result of clients increasingly exchanging data (users). As a result, it has lately come to light that cryptography systems' security level evaluation component is crucial. In this study, a hybrid ANFIS facility combining Artificial Neural Network (ANN) and Fuzzy Logic approaches is used in the editor of MATLAB (2021a). The encryption technique (Lattice-Based Crypto) has been designed and programmed using a simulation based on its original algorithm. The Lattice-Based Crypto technique is therefore used to acquire data values for the following parameters (message length, execution time, length of key, and cipher message entropy). They executed specialized software programs that selected values as the training sample number. The ANFIS editor was opened, the training data was loaded, the key ANFIS parameters were identified, and the training data was trained using the lowest error tolerance. The best membership function is empirically found to be assigned for the ‘Trimf’ for the Lattice-Based Cryptography. where it has achieved the best performance in the training phase. (Trimf=0.35355, Trapmf=0.41601, Gbellmf=0.48539, Gaussmf=0.43792, Gauss2mf=0.47423, Pimf=0.41197, Dsigmf=0.40179, Psigmf=0.6574).