Improved Column Generation Algorithms for the Job Grouping Problem

Dantzig-Wolfe reformulation solved by Column Generation is an approach to obtain improved bounds for Mixed Integer Programs. A downside of this approach is that the Column Generation process can be very time consuming due to degeneracy of the master problem and the tailing-off effect. To counter this problem in the Column Generation process, we also perform a number of Lagrangean Relaxation iterations to generate new columns after each time that we solve the master problem. This potentially reduces the number of times we have to solve the master problem. We explore this combination of Column Generation and Lagrangean Relaxation for the Job Grouping Problem. In our test set, we observe that the time spent to solve the master problem is decreased by a factor 4, while the total time spent on solving the subproblems remains approximately equal.

[1]  Abdullah Konak,et al.  Minimizing the number of tool switching instants in Flexible Manufacturing Systems , 2008 .

[2]  Zeger Degraeve,et al.  Optimal Integer Solutions to Industrial Cutting-Stock Problems: Part 2, Benchmark Results , 2003, INFORMS J. Comput..

[3]  M. Wagner,et al.  Generalized Linear Programming Solves the Dual , 1976 .

[4]  Jacques Desrosiers,et al.  Dual-Optimal Inequalities for Stabilized Column Generation , 2003, Oper. Res..

[5]  Arthur M. Geoffrion,et al.  Lagrangian Relaxation for Integer Programming , 2010, 50 Years of Integer Programming.

[6]  Jacques Desrosiers,et al.  Stabilized dynamic constraint aggregation for solving set partitioning problems , 2011, Eur. J. Oper. Res..

[7]  Yves Crama,et al.  Models for machine-part grouping in cellular manufacturing , 1996 .

[8]  Jacques Desrosiers,et al.  Selected Topics in Column Generation , 2002, Oper. Res..

[9]  Francisco Barahona,et al.  Plant location with minimum inventory , 1998, Math. Program..

[10]  Marshall L. Fisher,et al.  The Lagrangian Relaxation Method for Solving Integer Programming Problems , 2004, Manag. Sci..

[11]  Meltem Denizel.  Minimization of the Number of Tool Magazine Setups on Automated Machines: A Lagrangean Decomposition Approach , 2003, Oper. Res..

[12]  Zeger Degraeve,et al.  BIBLIOGRAPHIC DATA AND CLASSIFICATIONS , 2003 .

[13]  Christopher S. Tang,et al.  Models Arising from a Flexible Manufacturing Machine, Part II: Minimization of the Number of Switching Instants , 1988, Oper. Res..

[14]  Jacques Desrosiers,et al.  A Unified Framework for Deterministic Time Constrained Vehicle Routing and Crew Scheduling Problems , 1998 .

[15]  Yves Crama,et al.  A column generation approach to job grouping for flexible manufacturing systems , 1994 .

[16]  Frits C. R. Spieksma,et al.  Production planning problems in printed circuit board assembly , 2002, Discret. Appl. Math..