针对行政事业单位全面预算管理中因过度依赖初始数据而陷入局部最优的问题,提出了基于企业资源计划(Enterprise Resource Planning,ERP)系统的行政事业单位全面预算管理法(Comprehensive budget management law of administrative institutions,CBMLAI)。利用ERP系统的随机性和全局搜索性找到指定数量的初始数据。在行政事业单位全面预算管理中,对传统的全面预算管理系统进行优化,以避免CBMLAI陷入局部最优,进而获得精度更高的解。为了验证该算法的性能,将CBMLAI用于不同规模的标准行政事业单位数据集中,实验结果表明,CBMLAI与聚类(K-means)数据融合算法(Data fusion algorithm,DFA)对比显示更好的数据性能和更好的稳定性和鲁棒性,CBMLAI在全面预算管理方面能取得更好的效果。

In view of the issue of local optimum caused by too much dependence on primary data in comprehensive budget management of administrative institutions,this puts forward a comprehensive budget management law of administrative institutions( CBMLAI) based on enterprise resource planning( ERP) system. The specified amount of primary data can be found by taking the advantage of the randomness and global searching performance of ERP system. In comprehensive budget management of administrative institutions,it is proposed to optimize the traditional comprehensive budget management system to avoid the issue of local optimum so as to get more accurate solutions. In order to verify the performance of this algorithm,CBMLAI is applied in the standard data concentrations in different scales of administrative institutions. The result shows that,compared with K-means and data fusion algorithm( DFA),CBMLAI has better data performance,steadiness and robustness. So,CBMLAI can help to achieve better effect in comprehensive budget management.