遗传算法与变邻域搜索混合模型在护士排班中的应用  

The hybrid model of genetic algorithm and variable neighborhood search for nurse scheduling problem

作　　者：孙雪松[1] 郑西川[1] 郝安琪 陈霆[1] 

机构地区：[1]上海交通大学附属第六人民医院,上海200233

出　　处：《北京生物医学工程》2015年第6期612-615,共4页Beijing Biomedical Engineering

基　　金：上海市信息化发展专项基金(201101049)资助

摘　　要：目的在探讨基于层级管理的护士APN排班模型的基础上,引入具有精确的全局搜索能力的遗传算法与变邻域搜索混合模型来解决护士排班过程中人员层级复杂、不同班次人员需求量不同等问题。方法首先,初始化排班表,然后利用遗传算法对初始化的排班表进行搜索,得到一个初步的排班表,最后利用变邻域搜索优化排班表,得到满足要求的排班表。结果使用该混合模型得到的排班表,能满足每日护士的人数需求和每名护士的工作量要求。同时,基于层级管理,各层级护士分开排班,能降低排班难度,在增加护士人员时,效率仍然较高。结论遗传算法与变邻域搜索的混合模型能解决护士排班中的各种复杂问题,是一种符合实际需求的模型。基于此模型来编制护士排班软件,能提升护理工作效率和满意度,便于对护理人员进行量化考核统计分析。Objective To introduce genetic algorithm （GA） ＆ variable neighborhood search （VNS） for solving the complex problems appearing in the scheduling process such as the different level of staff, the different number of nurses for different shift when exploring the APN scheduling model based on the nurse-level management. Methods First, we initialize the scheduling table. Second, GA is used to explore the scheduling table and another table is obtained. Third, VNS is used to improve the table and the final scheduling table is obtained. Results The hybrid model can satisfy the complex requirements. Moreover, the exploring process is based on the different level of nurses which is still effective for more nurses. Conclusions The hybrid model satisfies the nurse scheduling requirements and can be used for encoding to realize scheduling process based on the software.

关 键 词：APN排班模式 遗传算法 变邻域搜索 混合模型 

分 类 号：R318[医药卫生—生物医学工程]