A Genetic Surrogate Model-Based Optimization Method for Reducing the High Computational Complexity of GEP-ECOC

With the development of wireless communication systems, the complexity of channels and the requirements for data processing rates have gradually increased. Therefore, designing efficient and highly stable error correction codes (ECC) has become indispensable to communication systems. Gene Expression Programming-Error Correction Output Code (GEP-ECOC) is a dynamic error correction code that integrates error correction output codes with Gene Expression Programming, enabling adaptability to dynamic and highly complex channel environments. However, optimizing GEP-ECOC matrices is inherently a Non-deterministic Polynomial-time Hard problem (NP-hard), resulting in high computational complexity. This paper aims to reduce the computational complexity of GEP-ECOC by constructing a genetic surrogate optimization model (GSOM) to replace the traditionally high-complexity expression tree. Additionally, a simplified GEP-ECOC model and a GSOM are implemented in MATLAB. A low-complexity GEP-ECOC coding matrix generation algorithm is designed, and both models are evaluated through comparative experiments in a conventional additive white Gaussian noise (AWGN) wireless communication channel. The results demonstrate that while slightly increasing the bit error rate (BER), the proposed approach significantly improves computational efficiency, enhancing the practical applicability of GEP-ECOC coding in conventional wireless communication systems.