Integrated Operational Planning and Medium-Term Scheduling for Large-Scale Industrial Semicontinuous/Continuous Processes

The operational planning and the medium-term scheduling of industrial semicontinuous/continuous multiproduct processes are closely linked to each other, since both operational planning and scheduling involve the allocation of plant resources. The operational planning problem typically determines production targets for the plant to meet product demands over a long time horizon from one to three months. The medium-term scheduling problem is defined over a shorter time horizon of between two to four weeks and involves detailed decisions such as the sequence of processing units, and detailed timings in which various products should be processed in each unit to meet the production targets supplied by the operational planning problem [1-2]. The lack of integration of planning and scheduling often make the planning model to provide unrealistic production targets causing the scheduling model to allocate plant resources in a suboptimal manner. The effective integration of planning and scheduling can increase profits and reduce committed capital simultaneously [3]. However, because of their disparate time scales, the effective integration of planning and scheduling is a critically challenging task. Therefore, it is imperative to propose a novel framework that can integrate the operational planning and scheduling more effectively.

A simple approach for solving the integration of planning and scheduling problems is to develop a single simultaneous planning and scheduling model over the entire the planning horizon. However, this approach resulted in large problem size for typical planning horizons, which is often computationally intractable. Therefore, the integration problem is often solved using bilevel decomposition approach [4], and rolling horizon algorithm [5]. Shah [6], Kallrath [7], Maravelias and Sung [8], and Verderame et al. [9] presented excellent reviews on the integration of planning and scheduling. However, most of the work related to integration of planning and scheduling has focused on batch processes. Moreover, most of the proposed operational planning models provide the aggregate production targets, not the daily production profile required by the scheduling model.

Recently, Verderame and Floudas [10-15] developed novel discrete-time mixed-integer linear programming (MILP) operational planning with production disaggregation models (PPDM) for a large-scale multipurpose and multiproduct batch plant with/without uncertainty. In their model, they disaggregated the production totals into a feasible distribution of daily production requirements and hence provided the daily production profile. They also allowed any unutilized reactor time to carry over from one day to the next and hence captured the continuous time nature of the plant. To integrate the operational planning and medium-term scheduling model developed by Janak et al. [16-17] effectively, they proposed a novel forward rolling horizon framework that allowed the feedback of scheduling production leading to additional operational planning model constraints. The computational results showed that the proposed framework yields greater aggregate production totals and a higher degree of daily demand satisfaction, compared to the approach of isolated planning and scheduling, which did not allow the two-way interaction between planning and scheduling. However, all of their work is also limited to batch processes.

In this presentation, we first introduce a novel mixed-integer linear programming (MILP) operational planning model based on discrete-time representation for large-scale industrial semicontinuous/continuous multiproduct processes. In the model, we allow unused processing time to carry over from one day to the next and capture the continuous time nature of the plant. The production totals are disaggregated into a feasible distribution of daily production requirements and hence daily production profiles are provided to meet the requirements of medium-term scheduling model developed by Shaik and Floudas [18]. The forward rolling horizon framework proposed by Verderame and Floudas [10-15] is utilized to effectively address the integration of operational planning and medium-term scheduling model. The forward rolling horizon framework is applied to an industrial case study of a large-scale continuous multiproduct plant over a three-month time horizon. The computational results show that the framework can lead to the operational planning model providing a tighter upper bound on the production capacity and a higher degree of daily demand satisfaction compared to the approach of isolated planning and scheduling.

References

[1] Floudas CA, Lin X. Continuous-time versus discrete-time approaches for scheduling of chemical processes: A review. Computers and Chemical Engineering. 2004, 28, 2109-2129.

[2] Floudas CA, Lin X. Mixed integer linear programming in process scheduling: Modeling, algorithms, and applications. Annals of Operations Research. 2005, 139, 131-162.

[3] Shobrys DE, White DE. Planning, scheduling and control systems: Why cannot they work together. Computers and Chemical Engineering. 2002, 26, 149-160.

[4] Erdirik-Dogan M, Grossman IE. A decomposition method for the simultaneous planning and scheduling of single-stage continuous multiproduct plants. Industrial and Engineering Chemistry Research. 2006, 45, 299-315.

[5] Li ZK, Ierapetritou MG. Rolling horizon based planning and scheduling integration with production capacity consideration. Chemical Engineering Science. 2010, 65, 5887-5900.

[6] Shah N. Process industry supply chains: Advances and challenges. Computers and Chemical Engineering. 2005, 29, 1225-1235.

[7] Kallrath J. Planning and scheduling in the process industry. OR Spectrum. 2002, 24, 219-250.

[8] Maravelias CT, Sung C. Integration of production planning and scheduling: Overview, challenges and opportunities. Computers and Chemical Engineering. 2009, 33, 1919-1930.

[9] Verderame PM, Elia JA, Li J, Floudas CA. Planing and scheduling under uncertainty: A review across multiple sections. Industrial and Engineering Chemistry Research. 2010, 49, 3993-4017.

[10] Verderame PM, Floudas CA. Integrated operational planning and medium-term scheduling for large-scale industrial batch plants. Industrial and Engineering Chemistry Research. 2008, 47, 4845-4860.

[11] Verderame PM, Floudas CA. Operational planning framework for multisite production and distribution networks. Computers and Chemical Engineering. 2009, 33, 1036-1050.

[12] Verderame PM, Floudas CA. Operational planning of large-scale industrial batch plants under demand due date and amount uncertainty. I. Robust optimization frameowrk. Industrial and Engineering Chemistry Research. 2009, 48, 7214-7231.

[13] Verderame PM, Floudas CA. Integration of operational planning and medium-term scheduling for large-scale industrial batch plants under demand and processing time uncertainty. Industrial and Engineering Chemistry Research. 2010, 49, 4948-4965.

[14] Verderame PM, Floudas CA. Operational planning of large-scale industrial batch plants under demand due date and amount uncertainty: II. Conditional value-at-risk framework. Industrial and Engineering Chemistry Research. 2010, 49, 260-275.

[15] Verderame PM, Floudas CA. Multisite planning under demand and transportation uncertainty: Robust optimization and conditional value at risk framework. Industrial and Engineering Chemistry Research. In press, 2011. DOI: 10.1021/ie101401k.

[16] Janak SL, Floudas CA, Vormbrock N. Production scheduling of a large-scale industrial batch plant. I. Short-term and medium-term scheduling. Industrial and Engineering Chemistry Research. 2006, 25, 8234-8252.

[17] Janak SL, Floudas CA, Vormbrock N. Production scheduling of a large-scale industrial batch plant. II. Reactive scheduling. Industrial and Engineering Chemistry Research. 2006, 45, 8253-8269.

[18] Shaik MA, Floudas CA, Kallrath J, Pitz HJ. Production scheduling of a large-scale industrial continuous-plant: Short-term and medium-term scheduling. Computers and Chemical Engineering. 2009, 33, 670-686.