基于改进BP算法的企业技术专家资源任务配比评价研究  

Research on the evaluation of resource task ratio of enterprise technical experts based on improved BP algorithm

作　　者：李立 王川 LI Li;WANG Chuan(Shanghai Aurora College,Shanghai 201900,China;Hangzhou Dianzi University,Hangzhou 310018,China)

机构地区：[1]上海震旦职业学院,上海201900 [2]杭州电子科技大学,浙江杭州310018

出　　处：《粘接》2024年第10期125-128,共4页Adhesion

基　　金：教育部产学合作协同课题(项目编号:21BY32567149)。

摘　　要：为有效提高专家库资源规划评价系统中任务分配效率,利用神经网络算法构建配对比较矩阵,并与随机森林算法进行比较。仿真结果表明,当任务数量为50时,神经网络算法的分配率高达84%,而随机森林算法的分配率仅为60%。专家数量为75时,神经网络算法的分配率高达73%,而随机森林算法的分配率仅为59%,较神经网络算法降低19.18%。当阈值g为1时,神经网络算法进入分配状态的任务数量多,且分配成功率高达90%。且神经网络算法平均分配时间最短为1439.9 ms,最长时间为4905.7 ms,而随机森林算法平均分配时间最短为2047.1 ms,最长时间为7219.1ms。In order to effectively improve the efficiency of task allocation in the expert database resource planning and evaluation system,a neural network algorithm was used to construct a pairing comparison matrix,and compare it with the random forest algorithm.The simulation results showed that when the number of tasks was 50,the allocation rate of the neural network algorithm was as high as 84%,while the allocation rate of the random forest algorithm was only 60%.When the number of experts was 75,the allocation rate of the neural network algorithm was as high as 73%,while the allocation rate of the random forest algorithm was only 59%,which was 19.18%lower than the neural network algorithm.When the threshold g was 1,the neural network algorithm entered the allocation state with a large number of tasks and a success rate of up to 90%.The average allocation time of the neural network algorithm was 1439.9 ms,and the longest time was 4905.7 ms,while the average allocation time of the random forest al⁃gorithm was 2047.1 ms,and the longest time was 7219.1 ms.

关 键 词：神经网络算法 化工企业专家库 任务 分配成功率 分配时间 

分 类 号：TP391.9[自动化与计算机技术—计算机应用技术]