The development of powerful computational tools in the last decades has introduced new approaches to chemical process synthesis. Coupled with advances in numerical techniques for optimization, process design is a significant area of research for many industrial companies. These tools have been used extensively for the development of economically optimal processing plants. The analysis for process synthesis has focused in most cases on identifying economically optimal plants, by determining the conditions that result in maximum conversion, minimum investment, or similar performance criteria.

The application of these methods can result in economically optimal operations, but other important areas such as the dynamic behavior of the system and the safety of the process have to be analyzed as well, as economically optimal processes may be unstable or unsafe. A bilevel programming approach is presented for ensuring a stable and economically optimal, or safe and economically optimal processing plant. This method relies on inter-nested optimization problems, where one of the problems is a constraint for the other one.

Introducing two novel solution techniques to these types of optimizations, this thesis presents a stochastic method, based on simulated annealing, to optimize a plant layout problem, where the cost minimization includes the cost of a worst-case scenario. Further, an alternative, deterministic technique, is used design chemical reactor networks that maximize conversion, while ensuring stability in case fluctuations in feed conditions occur. These networks are developed from a steady state operational focus or a dynamic stability perspective.