Planning and scheduling of batch processes under uncertainty

The thesis centers upon the application of mathematical modeling, uncertainty analysis, and optimization theory and algorithms to the problem of planning and scheduling of batch processes under uncertainty. Operational planning, medium-term scheduling, and short-term scheduling are closely related activities; however, despite prior advances, the effective integration of planning and scheduling remains a challenging problem, especially for large-scale industrial processes. In order to address this issue, work has been conducted on the integration of planning and scheduling and multisite planning. Novel applications of robust optimization and conditional value-at-risk theory have been utilized in order to explicitly take into account all pertinent forms of parameter uncertainty at both the planning and scheduling level.;The operational planning and the medium-term scheduling of a multipurpose and multiproduct batch chemical plant are interrelated activities dealing with the allocation of plant resources. Due to their disparate time scales, however, the effective integration of planning and scheduling is a formidable task. The lack of an integrative framework for planning and scheduling will invariably cause the planning model to provide unrealistic production targets to the scheduling level leading to the misallocation of plant resources. In response to this issue, a novel framework for the integration of planning and scheduling for a multipurpose and multiproduct batch plant has been developed. The framework entails integrating a novel Planning with Production Disaggregation Model with a medium-term scheduling model through a forward rolling horizon approach. Both robust optimization and conditional value-at-risk theory have been utilized in order to take into account demand due date and amount uncertainty at the planning level and processing time uncertainty at the scheduling level.;The operational planning of a multisite production and distribution network, which entails making operational level decisions for efficient production facility utilization and customer-order fulfillment over a time horizon of several months, is of great importance but has received considerably less attention compared to operational planning approaches for a single production site. The inherent complexities of simultaneously optimizing the allocation of production tasks at each facility, as well as the interplay between several production and distribution centers make operational planning of a multisite production and distribution network challenging especially when addressing large-scale industrial applications. The proposed Multisite Planning with Production Disaggregation Model has been formulated in order to address industrially relevant supply chains and determine both the production and shipment profiles for the supply chain of interest. Robust optimization, as well as conditional value-at-risk theory have been applied in order to explicitly model demand due date and amount uncertainty and transportation time uncertainty.