Two-stage stochastic days-off scheduling of multi-skilled analysts with training options

作     者：Altner, Douglas S. Mason, Erica K. Servi, Les D. 

作者机构：Mitre Corp Operat Res Dept Mclean VA 20111 USA ThirdLove San Francisco CA 94107 USA 

出 版 物：《JOURNAL OF COMBINATORIAL OPTIMIZATION》 (组合优化杂志)

年 卷 期：2019年第38卷第1期

页      面：111-129页

核心收录：

学科分类：12[管理学] 120202[管理学-企业管理（含：财务管理、市场营销、人力资源管理）] 0202[经济学-应用经济学] 02[经济学] 1202[管理学-工商管理] 1201[管理学-管理科学与工程(可授管理学、工学学位)] 0701[理学-数学] 0812[工学-计算机科学与技术（可授工学、理学学位）] 

基　　金：We would like to thank our MITRE co-worker Kael Stilp for his comments on an earlier draft of this manuscript  which have helped us improve the exposition and content of the paper. We would also like to thank comments from our anonymous reviewers  who we think helped improved the clarity and exposition of this paper and brought a few interesting studies to our attention 

主　　题：Days-off scheduling Multi-skilled workforce Stochastic integer programming Matheuristics Cybersecurity operations Training 

摘      要：Motivated by a cybersecurity application, this paper studies a two-stage, stochastic days-off scheduling problem with (1) many types of jobs that require specialized training, (2) many multi-skilled analysts, (3) the ability to shape analyst skill sets through training decisions, and (4) a large number of possible future demand scenarios. We provide an integer linear program for this problem and show it can be solved with a direct feed into Gurobi with as many as 50 employees, 6 job types, and 50 demand scenarios per day without any decomposition techniques. In addition, we develop a matheuristicthat is, an integer-programming-based local search heuristicfor instances that are too large for a straightforward feed into a commercial solver. Computational results show our matheuristic can, on average, produce solutions within 4-7% of an upper bound of the optimal objective value.