Homepage        Research        Pictures        Istanbul        Links        Contact

A HEURISTIC APPROACH FOR BIG-BUCKET PRODUCTION PLANNING PROBLEMS

Akartunali, K. and A.J. Miller, "A Heuristic Approach for Big Bucket Multi-Level Production Planning Problems", European Journal of Operational Research, 193: 396-411, 2009.     PDF        

This page is designed to provide researchers all the lot-sizing test sets we experimented with in our paper, the paper itself and sample mosel files with our heuristic for test purposes (please check first the read-me files).  Please feel free to use these, and let us know if you encounter any problems.

TEST SETS:

1) Test instances generated by [4] and [3]: These include overtime variables. Sets A+ and B+ involve problems with 10 items, 24 periods and 3 machines, and sets C and D involve problems with 40 items, 16 periods and 6 machines. Sets B+ and D include setup times. We chose the hardest instances of each data set for our computations, i.e., for each data set, we picked 10 assembly and 10 general instances with the highest duality gaps according to Stadtler’s [3] results.

Click here for the official website with more information and data files

.zip file with sample mosel files and data files for a test instance 

2) LOTSIZELIB [1]: We tested the multi-level big-bucket models of this library. These include single-machine problems with big bucket capacities. Backlogging is allowed. The problems vary between 40 item, single end-item problems and 15 item, 3 end-item problems, with both assembly and general BOM structures. All problems have 12 periods.

Click here for the official website with more information and data files

.zip file with sample mosel files and data files for a test instance

3) MULTILSB: We have generated 4 sets of test problems based on the problem described in the paper of Simpson and Erenguc [2], each set having 30 instances with low, medium and high variability of demand. These instances are different from the previous problems in that they take component commonality into consideration and hence consider setup variables for each family so that setup times are defined for each family of items instead of for each item. Except the problems in SET2, which considers a horizon of 24 periods, all the instances have 16 periods. All instances consider 78 items and have an assembly structure, and all instances allow backlogging in the last period.

Click here for the official website with more information and data files

.zip file with sample mosel files and data files for a test instance

[2]: N.C. Simpson and S.S. Erenguc. Modeling multiple stage manufacturing systems with generalized costs and capacity issues. Naval Research Logistics, 52:560–570, 2005.

[3]: H. Stadtler. Multilevel lot sizing with setup times and multiple constrained resources: Internally rolling schedules with lot-sizing windows. Operations Research, 51:487–502, 2003.

[4]: H. Tempelmeier and M. Derstroff. A lagrangean-based heuristic for dynamic multilevel multiitem constrained lotsizing with setup times. Management Science, 42(5):738–757, 1996.