基于遗传算法优化的BP神经网络的禽病辅助诊断与管理信息系统的设计与实现

Design and implementation of poultry disease diagnosis and information management system based on BP Neural Network optimized by genetic algorithm

为了满足基层养殖场对兽医服务的需要,降低养殖场家禽发病率,研究采用B/S结构模式设计遗传算法优化的BP神经网络的禽病辅助诊断与管理信息系统,应用Visual Basic.NET、Asp.NET开发语言与SQL Server 2008数据库等实验工具进行构建,采用遗传算法优化BP神经网络的推理机制进行疾病诊断专家系统的开发,运用遗传算法在全局空间上搜索权值和阈值的最优解,再用BP算法在最优点附近寻求最优值。结果表明:在实际应用中,系统运行稳定,误诊率低,可以提高养殖场管理效率,可用于进一步分析与挖掘养殖场生产数据中的潜在信息,提高利用效率。

To meet the needs of the grassroots farms for veterinary services, and reduce the incidence of poultry diseases in poultry farms, B/S structure mode was adopted to design poultry disease diagnosis and information management system based on BP Neural Network optimized by genetic algorithm in the study. The system was constructed using Visual Basic. NET, Asp. NET development language and SQL Server 2005 da- tabase experimental tools, and then the inference mechanism of BP Neural Network optimized by genetic algorithm was used to develop disease diagnosis expert system. Furthermore, the genetic algorithm was used to search the optimal solution of the weights and thresholds in global space, and then the BP algorithm was used to seek the optimal value near optimal point. The results showed that in practical application, the system was stable in operation and had low misdiagnosis rate. The results indicate that the system can not only improve the efficiency of farm management but also can be used to further analyze and excavate the potential information in the farm production data to improve the utilization efficiency.