基于PSO-RF的妊娠母猪日饲喂量预测算法研究  

Research on the Prediction Algorithm of Daily Feeding Amount for Pregnant Sows Based on PSO-RF

作　　者：凌丽 樊晓宇 岳宝昌 谭飞飞 胡俊泽 任国栋 LING Li;FAN Xiaoyu;YUE Baochang;TAN Feifei;HU Junze;REN Guodong(School of Mechanical Engineering,Anhui Science and Technology University,Fengyang 233100,China;School of Electrical and Electronic Engineering,Anhui Science and Technology University,Bengbu 233030,China;Bengbu Yi Ai Electronic Technology Limited Liability Company,Bengbu 233030,China)

机构地区：[1]安徽科技学院机械工程学院,安徽凤阳233100 [2]安徽科技学院电气与电子工程学院,安徽蚌埠233030 [3]蚌埠依爱电子科技有限责任公司,安徽蚌埠233030

出　　处：《内蒙古民族大学学报(自然科学版)》2025年第1期44-52,共9页Journal of Inner Mongolia Minzu University:Natural Sciences Edition

基　　金：安徽省高校自然科学研究重点项目(2022AH051633);安徽省农业物质技术装备领域揭榜挂帅项目(S202320230906020001);蚌埠市科技计划项目(2022gx10);安徽科技学院人才稳定项目(HCWD202001);安徽科技学院科研发展基金项目(FZ220116)。

摘　　要：日饲喂量对妊娠期母猪繁殖性能具有较大影响,对于保障母猪健康、胎儿生长发育具有重要意义。为了精准控制日饲喂量,针对粒子群算法(PSO)各阶段搜索能力不均衡的问题,引入一种非线性递减惯性权重策略对PSO算法进行改进,并用改进的粒子群优化随机森林回归算法(PSO-RF)精确预测妊娠母猪日饲喂量,精准控制智能饲喂器的饲料投放。该算法融合随机森林的高准确性和粒子群算法的参数寻优能力强的特性,通过优化决策树的数量和最大深度来提升预测性能。结果表明,PSO-RF算法取得的决定系数R^(2)值达到0.9814,相较于RF算法、SVM支持向量机和BP神经网络分别提升了1.19%、2.30%和3.25%。PSO-RF算法在预测妊娠母猪日饲喂量方面具有更高的精准度,有助于提高养猪场管理的智能化水平,降低生产成本,提升养猪场养殖效益,具有一定实际应用价值。The daily feeding amount has a great impact on the reproductive performance of pregnant sows and is of great significance for ensuring the health of sows and the growth and development of fetuses.In view of the unbalanced search ability of Particle Swarm Optimization(PSO)algorithm,a nonlinear decreasing inertial weight strategy was introduced to improve the PSO algorithm,and the improved particle swarm optimization random forest regression algorithm(PSO-RF)was used to accurately predict the daily feeding amount of pregnant sows,and accurately control the feed delivery of intelligent feeding device.The algorithm combines the high accuracy of the Random Forest algorithm with the strong parameter finding ability of the Particle Swarm Optimization(PSO)algorithm to enhance the prediction performance by optimizing the number and maximum depth of decision trees.The results demonstrate that the PSO-RF algorithm attains a coefficient of determination R^(2)value of 0.9814,representing an enhancement of 1.19%,2.30%,and 3.25%in comparison to the RF algorithm,SVM(Support Vector Machine),and BP(Neural Network),respectively.The PSO-RF algorithm demonstrates superior accuracy in predicting the daily feeding amount of pregnant sows,which can facilitate enhancements in the intelligence of pig farm management,reduce production costs and improve the farm’s breeding efficiency.Consequently,it possesses definite practical application value.

关 键 词：妊娠母猪 日饲喂量 随机森林回归算法 粒子群优化算法 PSO-RF 

分 类 号：S23[农业科学—农业机械化工程]