Implementation of the State-Task-Network (STN) [Kondili et al. 1993] with degradation of equipment. This code can be used to replicate the results in [Wiebe et al. 2018] available at arXiv:1810.09289.

This implementation is based on the STN-Scheduler by Jeffrey Kantor (c) 2017.

• Pyomo
• A MILP solver. The module has been tested with CPLEX.

Generate data for logistic regression by solving short-term scheduling model repeatedly for different demands:

python lhs.py runs/test_lhs.yaml

where runs/test_lhs.yaml is a config file.

Train logistic regression for Markov-chain or frequency approach.

Solve model using rolling horizon:

python rolling runs/test_det.yaml

Solve model using Markov-chain or frequency approach:

python mc.py runs/test_mc.yaml

Optimize uncertainty set size using Bayesian Optimization:

python bo.py runs/test_bo.yaml prefix_for_file_names

Kondili, E.; Pantelides, C.; Sargent, R. A general algorithm for short-term scheduling of batch operations - I. MILP formulation. Computers & Chemical Engineering 1993, 17, 211227.

Wiebe, J.; Cecilio, I.; Misener, R. Robust optimization of processes with degrading equipment. arXiv:1810.09289. 2018 (Accepted).