(298f) Rolling Horizon Based Planning and Scheduling Integration with Production Capacity Consideration

Li, Z. - Presenter, Rutgers University
Ierapetritou, M. G. - Presenter, Rutgers, The State University of New Jersey

To ensure the consistency between production planning and scheduling, integrated decision making is necessary. Among the various types of aggregation, relaxation and decomposition methods, rolling horizon method has received a lot of attention not only in the literature but also in practical industrial applications [1], [2]. The main advantage of the rolling horizon methodology is that it is computationally efficient and can lead to feasible planning-scheduling decisions. However, a major problem of most existing rolling horizon based methods is that they often rely on the simplistic representation of the scheduling problem and thus the quality (optimality) of the solution cannot be ensured.

In this work, we first study the effect on the final solution's quality by including production capacity information within the planning problem of the rolling horizon method for planning and scheduling integration. It is found that the solution quality can be significantly improved through the incorporation of production capacity information into the planning model. This is mainly due to the fact that the planning model can "predict" the future cost due to possible backorders with the production capacity information. However, the consideration of production capacity in planning model has not received a lot of attention in the literature with the exception of few studies (computational geometry based method [3]. Thus there is a need for the development of accurate production capacity information. In this work, a parametric programming based method [4] is proposed to address this issue. The proposed method solves the parametric optimization problem by considering the production targets of certain products as parameters and other production targets as objectives, thus the method can generate the production capacity information based on the short-term scheduling optimization problem. To avoid the large computational complexity of parametric programming method for high dimensionality problems (large number of products), a heuristic process network decomposition strategy is also proposed, so that the parametric programming method can be applied on the sub-networks which involve relative small number of products. The production capacity information is composed by the parametric solution derived within a sub-network and the parametric solution derived between sub-networks. The solution can be considered as the projection of the exact production feasibility space onto certain dimensions of the original feasible production space, and thus provides valid overestimation of the original feasible production capacity region.

A number of case studies have been studied to prove that the proposed method can improve the quality of the final solution comparing to the simple production capacity constraints derived by production recipe and mass balance. Finally, it is worth to point out that the proposed solution framework can be also applied to address the long-term and mid-term scheduling problem within a rolling horizon framework.

[1] Verderame, P.M., Floudas, C.A., 2008. Integrated operational planning and medium-term scheduling of a large-scale industrial batch plants. Ind. Eng. Chem. Res., 47, 4845-4860.

[2] Kreipl, S., Pinedo, M., 2004. Planning and Scheduling in Supply Chains: An Overview of Issues in Practice. Production and Operations Management, 13, 77-92.

[3] Sung, C., & Maravelias, C. T., 2007. An attainable region approach for effective production planning of multi-product processes. AIChE Journal, 53, 1298-1315.

[4] Li, Z., Ierapetritou, M.G., 2007. Process scheduling under uncertainty using multiparametric programming. AIChE Journal, 53, 3183-3203.