Integration of planning, scheduling, and control: A review and new perspectives

Abstract

The competitive, profitable, and safe operation of chemical plants depends on tight and effective coordination among the different decision making levels of the enterprise, including planning, scheduling, and control. The optimal integration of these functions has become critical given the disruptive effects of the recent COVID-19 pandemic on the supply chains and the current trends in climate change. However, integrating multiple decision making levels creates modelling and computational challenges. In this study, we review the progress made in the integration of two and three decisions levels using mathematical programming and control theory tools. We highlight the model reduction and decomposition techniques that have been applied, as well as the main issues that remain unsolved. Perspectives on emergent areas of application and novel computing solutions are also discussed.

A Benders decomposition framework for the optimization of disjunctive superstructures with ordered discrete decisions

“…Alternatively, external variables for which f à is known prior starting the algorithm can be added to D init ; this can be interpreted as a warm start of the algorithm. 24 After its initialization, D k will be recursively defined as indicated in Equation (10). If the subproblem (3) at iteration k is feasible (i.e., 11) and (12).…”

“…Some examples in the field of Process System Engineering include optimal process synthesis through superstructure optimization, 1,2 optimal scheduling, 3,4 problems involving the interaction of different operational decision layers such as optimal scheduling and control, [5][6][7][8][9] and optimal planning, scheduling and control. 10 Due to the development of rigorous multiscale models and frameworks that integrate multiple decision layers, problems are increasing in size and complexity; thus, algorithmic and theoretical developments are still in need to alleviate the resulting computational difficulties. 11 In particular, the literature suggests the exploitation of special properties and structures in the problems, for example, hidden convexity, for the development of tailored algorithms.…”

“…One common aspect of many optimization problems that arise in chemical engineering is the presence of nonlinearities and discrete decisions, which often result in complex nonconvex formulations. Some examples in the field of Process System Engineering include optimal process synthesis through superstructure optimization, 1,2 optimal scheduling, 3,4 problems involving the interaction of different operational decision layers such as optimal scheduling and control, 5–9 and optimal planning, scheduling and control 10 . Due to the development of rigorous multiscale models and frameworks that integrate multiple decision layers, problems are increasing in size and complexity; thus, algorithmic and theoretical developments are still in need to alleviate the resulting computational difficulties 11 .…”

Integration of Planning and Scheduling for Large-Scale Multijob Multitasking Batch Plants

“…In this approach, the planning and scheduling models are integrated by providing the scheduling model with the production targets obtained from the planning model with the goal of obtaining a scheduling solution. − Due to the one-way interaction framework, this approach restricts the ability to modify the planning problem with respect to the actual scheduling model components. The full space approach involves solving the planning and scheduling model as a single problem where detailed scheduling subproblems are used for each planning period. − This approach often requires the problem to be decomposable into master and subproblems as it utilizes decomposition/relaxation methods , to solve the otherwise computationally intractable problems. The iterative approach provides a two-way interaction between the planning and scheduling models .…”

Discrete-Time Network Scheduling and Dynamic Optimization of Batch Processes with Variable Processing Times through Discrete-Steepest Descent Optimization

“…The solution of this computationally demanding MINLP often requires advanced and specialized solution strategies; thus, research is still needed to develop reliable and efficient algorithms to increase the commercial adoption and implementation of integrated scheduling and control solutions. 3 While further research is needed in the online implementation of ISC solutions, this study presents research advances in the MINLP modeling and algorithmic solution of an optimization problem that enables the optimal integration of scheduling and control decisions. Thus, we consider an SSDO feedback strategy that generates an open-loop optimal schedule along with its corresponding open-loop control policies.…”