基于反向传播神经网络和遗传算法的新鲜Halloumi奶酪生产工艺优化

Production process optimization of fresh Halloumi cheese based on BP neural network and genetic algorithm

为提升Halloumi奶酪品质,采用反向传播神经网络和遗传算法优化Halloumi奶酪生产过程的多工艺参数。选取CaCl_(2)添加量、热烫温度和压榨压强为优化变量,以成品奶酪得率和感官评分为优化目标,分别建立了2个神经网络模型,模型精度分别达到了98.936%和98.255%。之后,通过遗传算法进行寻优,结果表明,在得率≥10%以及感官评分≥85的前提下,以奶酪得率为目标的最优生产工艺条件:CaCl_(2)添加量0.0144%、热烫温度83.5℃、压榨压强5.12 kPa,该条件下最高得率为12.01%。以感官品质为目标的最优生产工艺条件:CaCl_(2)添加量0.0171%、热烫温度83.7℃、压榨压强10.38 kPa,该条件下最高感官评分为94.5。该方法能够有效实现Halloumi奶酪生产工艺的快速优化,为促进Halloumi奶酪工业化提供理论基础。

To improve the quality of Halloumi cheese,BP neural network and genetic algorithm were used to optimize the multi-process parameters of Halloumi cheese production process.The addition amount of CaCl_(2),heating temperature,and pressing pressure were used as the optimization variables and two neural network models were established with the yield and sensory score of finished cheese as optimization objectives.The accuracies of the models reached 98.936% and 98.255%,respectively.After that,genetic algorithm was used to search for optimization.The results indicated that under the premise of yield rate higher than 10% and sensory score greater than 85,the optimal production conditions with cheese yield as target were determined as follows:CaCl_(2) addition of 0.0144%,heating temperature of 83.5℃,pressing pressure of 5.12 kPa,and the maximum yield rate reached 12.01%.The optimal production conditions for sensory quality were as follows:CaCl_(2) addition of 0.0171%,heating temperature of 83.7℃,pressing pressure of 10.38 kPa,and the highest sensory score reached 94.5.This method can effectively realize the rapid optimization of Halloumi cheese production process and provide a theoretical basis for promoting the industrialization of Halloumi cheese.