Coupling Optimization on the Fermentation Medium of Pinus Massoniana Resin by BP Neural Network and Genetic Algorithm

In order to improve the titer of Pinus massoniana resin degraded by the fermentation broth of Burkholderia sp.ZYB002, the medium was optimized. The single factor experiment was carried out to determine that 1% glucose was the most suitable carbon source for the fermentation, 0.3% urea and 2% corn flour were the most suitable compound nitrogen sources. The response surface method was used to obtain the optimal ratio of the three main control factors:glucose 0.576%, olive oil 1.81% and inoculation quantity 2.57%. After the optimization, the degradation titer of Pinus massoniana resin was increased to 42.5%. While the optimal ratio was obtained by the BP-GA coupled method:glucose 0.676 3%, olive oil 1.803 4%, inoculation quantity 3.381 3%. After the optimization, the degradation titer of Pinus massoniana resin was increased to 47.4%. The results also showed that compared with the response surface method, the BP-GA coupled method had the better optimization effect, and the titer after the optimization was improved by 38.6% compared with the initial degradation titer. After the optimization by the BP-GA coupled method, the optimal medium composition for the shake-flask fermentation of Burkholderia sp.ZYB002 was:glucose 0.676 3%, corn flour 1.2%, olive oil 1.803 4%, urea (nitrogen content) 0.05%, K₂HPO₄ 0.2%, NaHCO₃ 0.1%, Tween 80 1.0%, and initial pH 8.5. The culture condition was as follows:the fermentation temperature was 30℃, the inoculation quantity was 3.381 3%, the shaking speed was 220 r·min^-1, the loaded liquid was 35 mL(250 mL triangular flask), and the culture time was 36 h.