Aiming at the problem that large detection error and relies on initial weight and threshold in the detection of mixed gas using single BP neural network

a method of optimizing initial weights and thresholds of BP neural network by introducing cuckoo search（CS） algorithm with adaptive adjustment of discovery probability and simulated annealing algorithm is proposed to realized the quantitative detection of mixed gas.The core idea of the algorithm is used by the powerful global search ability of the cuckoo search algorithm to optimize the initial weight and threshold of the neural network.In the iterative process of the algorithm

the simulated annealing operation is utilized to avoid the algorithm falling into local optimum.Based on the introduction of the adaptive adjustment strategy of discovery probability for the new solution generation

the optimized parameters are used as the initial connection weights and thresholds for BP network to quantitatively identify the mixed gas.The experimental data are simulated on the MATLAB software.Compared with the previous single BP neural network algorithm

the improved algorithm has a significant quantitative detection effect on the mixed gas

and the average relative error is about 10%.