There has been a lot of attention in recent years towards the application of mathematical modeling and optimization approaches for the solution of production planning and scheduling problems. This is mainly due to the changing economic environment which pushes for more efficient process operations. However, there are still a number of challenges that restrict the effective application of optimization for planning and scheduling problem especially in the process industry. First, decision making in process operations is frequently based on parameters of which the values are uncertain. A systematic treatment of those uncertainties (e.g., processing time variations, rush orders, failed batches, machine breakdowns, etc) is necessary to satisfy the customer demands, increase the efficiency of operations and improve the plant profitability. Moreover, the interactions between the different decision-making levels were often ignored in existing solution approaches, which leads to sub-optimal and even infeasible solutions. Thus the integration of different decision making levels has been recognized by the research community as an imperative problem.

In this work, systematic methods have been developed to address the above challenges. First, different methodologies are proposed to address the uncertainties in process scheduling problem:​ robust optimization based preventive scheduling strategy which aims at generating a robust preventive schedule to handle the possible parameter perturbations;​ parametric programming based reactive scheduling strategy which aims at responsive schedule regeneration or updating upon the happening of disruptive events. To address the interaction between planning and scheduling decision levels, a dual decomposition based approach that targets the solution of large scale planning and scheduling integration problem was proposed, which aims at decreasing the computational complexity through decomposition and parallel computation. Finally, the rolling horizon method which provides a promising modeling framework for integrated planning and scheduling and incorporation of uncertainty is studied. A novel method of generating production capacity model through parametric programming technique is proposed, and it is verified that the incorporation of the capacity model into the rolling horizon framework can improve the solution quality.