New General Discrete-Time Scheduling Model for Multipurpose Batch Plants

Abstract

This work deals with the optimal short-term scheduling of general multipurpose batch plants, considering multiple operational characteristics such as sequence-dependent changeovers, temporary storage in the processing units, lots blending, and material flows traceability. A novel Mixed Integer Linear Programming (MILP) discrete-time formulation based on the State-Task Network (STN) is proposed, with new types of constraints for modeling changeovers and storage. We also propose some model extensions for addressing changeovers start; nonpreemptive lots; lots start and sizes; alternative task-unit and task-unitlayout assignments. Computational tests have shown that the proposed model is more effective than a similar model based on the Resource-Task Network (RTN).

Optimal planning and campaign scheduling of biopharmaceutical processes using a continuous-time formulation

“…The approaches based on a unified process representation, both the State-Task Network (STN) and the Resource-Task Network (RTN) proposed by Kondili et al (1993) and Pantelides (1994) respectively, have proven to be effective in most classes of scheduling problems. As an example of a real pharmaceutical industrial scheduling problem, Moniz et al (2013) proposed an MILP discrete time formulation based on the RTN framework, considering some production constraints, such as sequence-dependent changeovers, temporary storage in processing units, lots blending/splitting and materials traceability.…”

“…The RTN continuous-time formulation, as well as its discrete-time counterpart, considers binary N and continuous variables to characterise the event of task i starting at point t and ending at (or before) point t > t (Castro et al, 2004). Moreover, to assure typical regulatory policies of the pharmaceutical processes (Moniz et al, 2013), these variables are now extended to include a lot index l, congregate 4-indices iltt . The binary variable N iltt is equal to one if lot l of task i starts at event point t and finish until event point t , i,l,t,t gives the lot amount of material processed within the same time slot [t,t ].…”

“…According to Moniz et al (2013), a distinction must be made in the formulation of lots and task-batches. Lots characterise the amount of stable intermediate or final product produced throughout a known set of tasks executed in a known production sequence or recipe.…”

Optimal Scheduling of Multi-stage Multi-product Biopharmaceutical Processes Using a Continuoustime Formulation

“…As highlighted in the work developed by Moniz et al (2013), the chemicalpharmaceutical industry relies on critical production features concerning regulatory policies to assure production quality. The same is verified within the biopharmaceutical sector.…”

Changeover formulations for discrete-time mixed-integer programming scheduling models

“…If changeovers do not require resources and do not incur a cost, then changeover times can be enforced by simply allowing enough idle time between tasks (Kondili et al, 1993;Shah et al, 1993;Wolsey, 1997, Moniz et al, 2013. However, if resources are needed or costs need to be modeled, additional binary variables are necessary (Karmarkar and Schrage, 1985;Sahinidis and Grossmann, 1991;Kondili et al, 1993;Zentner et al, 1994;Kelly and Zyngier, 2007).…”