Systematic development of Constraint Programming formulations for industrial scheduling problems

Abstract

The problem of industrial scheduling is approached from a holistic perspective. This implies the construction of a scheduling system envisioned as a software product. In order to do that, a software development methodology is proposed and the most relevant contributions reached are synthesized. This work has been limited to modules and capabilities for supporting predictive scheduling problems in discrete manufacturing production environments. For the resolution of problems, constraint programming has been selected. An approach for automatic model generation in a systematic way is proposed, based on the analysis of the structure of the problems, the identification of decisions and restrictions, and their representation according to a proposed set of common modelling structures. A set of examples from the bibliography is used to assess whether the developed approach and the way models are generated lead to satisfactory results.