(614a) Operational Planning of a Large-Scale Industrial Batch Plant Under Uncertainty

The operational planning of a large-scale industrial batch plant typically occurs over a time horizon of several months with the goal of providing daily production targets and raw material requirements for the plant in question. Kallrath [1] and Shah [2] each provide an overview of the various aspects of planning. Due to the length of the time horizon, multiple forms of uncertainty should be taken into account in order to ensure that the operational planning model does not provide unrealizable production targets and/or raw material requirements. For example, demand, price, and processing time uncertainty should not be overlooked at the operational planning level. The main approaches for optimizing under uncertainty are stochastic programming, fuzzy programming, stochastic dynamic programming, and robust optimization, and the work of Lin et al. [3] and Janak et al. [4] has laid the foundation for a robust optimization approach toward addressing uncertainty within the field of planning.

Extending upon the work of Lin et al. [3] and Janak et al. [4], a novel operational planning model has been developed in order to address the objective of providing a reliable daily production profile (e.g., how much of each product should be produced at a facility on a daily basis) which is immune to various forms of uncertainty. The proposed formulation represents the robust counterpart of the novel Production with Demand Disaggregation Model (PPDM), a discrete-time mixed-integer linear programming (MILP) planning model developed by Verderame and Floudas [5] which still captures the continuous-time nature of the multipurpose and multiproduct batch plant under consideration. The given planning model is a unit aggregation model having the explicit objective of providing a daily production profile for a batch chemical plant. The PPDM has been extended in order to take into account uncertainty associated with the demand, price, and processing time parameters for both the bounded and known probability distribution cases. The ability of the robust PPDM to address the aforementioned objectives of an operational planning model has been validated through an industrial case study of a large-scale, multiproduct and multipurpose batch plant having the capability of producing hundreds of different products over a time horizon of three months.

[1] Kallrath, J. Planning and Scheduling in the Process Industry. OR Spectrum, 2002, 24, 219.

[2] Shah, N. Process Industry Supply Chains: Advances and Challenges. Comp. Chem. Eng. 2005, 29, 1225.

[3] Lin, X.; Janak, S.L.; Floudas, C.A. A new robust optimization approach for scheduling under uncertainty: I. Bounded uncertainty. Comp. Chem. Eng. 2004, 28, 1069.

[4] Janak, S.L.; Lin, X.; Floudas, C.A. A new robust optimization approach for scheduling under uncertainty: II. Uncertainty with known probability distribution. Comp. Chem. Eng. 2007, 31, 171.

[5] Verderame, P.M.; Floudas, CA. Integrated Operational Planning and Medium-term Scheduling of Large-Scale Industrial Batch Plants. Ind. Eng. Chem. Res. Accepted for Publication.