In this work we propose several reformulations of a general discrete-time mixed-integer formulation for production scheduling problems. Each reformulation involves the addition of new integer variables that we refer to as record keeping variables. Computational results show that the required CPU time is greatly reduced when record keeping variables are present in the optimization model. Additionally, we assess which of the individual aspects of traditional branch-and-bound solution procedures benefit from the inclusion of record keeping variables.
All Science Journal Classification (ASJC) codes
- Computer Science(all)
- Mixed integer programming
- Production scheduling
- Record keeping variables