General approach and tool for the scheduling of complex production systems

Abstract

The aim of this paper is to present a flexible modelling framework implemented into a basic general scheduling tool for both, the study of scheduling problems (for application developers and analysts) and as an efficient way to find practical solutions in complex industrial applications (for users). Thus, the basis to assist process modelling (simulation and simple optimisation procedures) is provided as well as an open framework capable to be easily customised (implementation of particular constraints, rules, exceptions, etc.). This approach and tool have been especially addressed to complex manufacturing systems (batch chemical processes, pharmaceutical, fine chemicals, etc.). Thus, the kind of recipes considered require a detailed description of tasks and subtasks and of the hard time and storage constraints implicated (simultaneous activities, unstable intermediates, product synthesis and other concurrent processes…). Towards this end, this work presents a general and practical description of production processes and recipes, as well as the mathematical model for the timing of the activities involved. Yet, the compact description attained does not introduce extra complexity to simpler cases. The approach introduced is based in a hierarchical description of the production processes and activities in which schedules are described by means of sequences of production runs associated to structured sets of assigned activities. This work shows how this sequences may be rearranged using general techniques as simulated annealing (SA), proving that this can be successful when specific rules are not available. Finally, different case studies are analysed to show the potential and flexibility of the approach presented.