(531d) A Web-Based Application for Chemical Production Scheduling

While a large range of models and methods have been proposed in the literature for chemical production scheduling, very limited tools are available for researchers to try and test these methods. In this work, we develop a web-based application to facilitate the generation and solution of chemical production scheduling problems. The application offers a graphical interface through which the user can define an instance using the state-task network representation (Kondili et al., 1993). The connections between the states and tasks can be created using connectors and processing data (e.g., task processing times, conversion coefficients) can be entered through the interface.

Given additional user-defined parameters (e.g., horizon, optimality gap) and choice of the objective function (e.g., profit maximization, cost minimization, makespan minimization), the application then automatically generates the corresponding mixed-integer linear programming discrete-time model. In terms of solution methods, the user can choose among three options: (1) tightening constraints, based on a constraint propagation algorithm (Velez et al., 2013); (2) reformulation based on the introduction of integer variables representing the total number of executed batches of each task (Velez and Maravelias, 2013); and (3) a discrete-continue algorithm which identifies good solutions fast based on a discrete-time model and then refines the solution based on a continuous time representation (Lee and Maravelias, 2018).

A data string representing the created network is generated automatically and submitted to the NEOS server (hosted at the University of Wisconsin-Madison). The user must provide their email address to retrieve the results. When the user submits the network through the web application, the NEOS server queues the relevant network data and based on the solution method(s) chosen, appropriate constraints are activated in the optimization model solved using the CPLEX solver in the GAMS platform. The user can access the model and solution statistics as well as a visual representation of the solution in the form of a Gantt chart along with the inventory profiles, which are zipped into a folder and emailed to the user.

References

1. Kondili, E., Pantelides, C.C., Sargent, R.W.H. (1993) A general algorithm for short-term scheduling of batch operations—I. MILP formulation. Computers & Chemical Engineering, 17, 211–227.
2. Velez, S., Sundaramoorthy, A., Maravelias, C.T. (2013) Valid inequalities based on demand propagation for chemical production scheduling MIP models. AIChE Journal, 59, 872–887.
3. Velez, S., Maravelias, C.T. (2013) Reformulations and branching methods for mixed-integer programming chemical production scheduling models. Industrial Engineering and Chemistry Research, 52, 3832–3841.
4. Lee, H., Maravelias, C.T. (2018) Combining the advantages of discrete and continuous-time scheduling models: Part 1. Framework and mathematical formulations. Computers & Chemical Engineering, 116, 176-190.