Research on Optimizing Electronic Nose Sensor Arrays for Oyster Cold Chain Detection Based on Multi-Algorithm Collaborative Optimization

Real-time quality monitoring during oyster cold chain transportation is a critical component in ensuring food safety. Addressing the issues of high redundancy and insufficient environmental adaptability in existing electronic nose systems, this study proposes a multi-algorithm collaborative optimization strategy for sensor array optimization. The system integrates ten gas sensors (TGS series, MQ series), employing Random Forest (RFA), Simulated Annealing (SA), and Genetic Quantum Particle Swarm Optimization (GA-QPSO) for sensor selection. KNN combined with K-means analysis validates the optimization outcomes. Under cold chain environments at 4 °C, 12 °C, 20 °C, and 28 °C, a multidimensional dataset was constructed by extracting global variables using feature correlation functions. Experiments demonstrate that the optimized sensor count decreases from 10 to 5-6 units while maintaining recognition accuracy above 95%, with redundancy decreased by over 40%. This multi-algorithm collaborative optimization effectively balances sensor array recognition precision, resource efficiency, and environmental adaptability, providing an intelligent, high-precision technical solution for oyster cold chain monitoring.