Exploiting the Multiscale Structure of Sustainable Engineering Problems via Network-Based Decomposition

In this chapter, we present the application of stochastic block modeling and Bayesian inference for analyzing and learning the structure of an optimization problem. We also show how the learned structure can be used as the basis for the application of distributed and hierarchical decomposition-based algorithms. This procedure is useful for complex optimization problems that emerge in the design and operation of sustainable processes and supply chains and require decomposition-based approaches for their solution. We discuss an example of transitioning an existing ammonia supply chain to a green ammonia one, and under different considerations, we highlight the structure of the resulting optimization problem and document how the proposed automatic structure detection framework can be used to enable efficient solution.