(371e) Supply Chains Optimization through Hybrid Tools

In recent years, there has been a significant upsurge in the utilization of supply chains within process engineering, driven primarily by the imperative to effectively meet specific resource demands across diverse regions. This upsurge has predominantly relied on mathematical programming, leveraging deterministic optimization models to ascertain the optimal configuration involving resources, equipment, suppliers, and final destinations.

However, a notable challenge encountered in the application of these optimization models lies in the inherent simplification entailed in designing production or generation equipment for resources intended for distribution through the proposed chains. This simplification is imperative to address the intricacies of the scenario. Recent advancements in chemical process engineering, encompassing design, simulation, and optimization, have substantially enhanced the treatment of supply chains. These advancements have yielded economic, environmental, and social benefits, including reduced production and operational costs, minimized raw material usage, decreased environmental emissions, and augmented job opportunities.

Traditionally, deterministic optimization has been employed to design these supply chains. Nonetheless, during the formulation of the mathematical model, numerous simplifications are made regarding the operation of the requisite equipment due to the presence of numerous equations containing highly nonlinear and nonconvex terms, posing challenges for deterministic solution methods. Historically, process units were often treated simplistically as black boxes in design, utilizing deterministic techniques for mathematical model solutions. However, the outcomes frequently diverged significantly from reality.

Consequently, hybrid strategies have been recently introduced, optimizing process units stochastically before conducting deterministic supply chain optimization. Although these strategies yield superior results, the sequential execution substantially escalates computational time. To address this issue, our work proposes a holistic solution that amalgamates metaheuristic methods with deterministic strategies to rigorously model equipment design, with the aim of developing an algorithm that effectively resolves these complexities.

By achieving synergy between deterministic and stochastic optimization software, our methodology utilizes data exchange and Visual Basic-programmed subroutines. We presented two distinct distillation sequences, with direct distillation yielding superior results compared to the indirect method, showcasing lower values for both economic and environmental objectives. These findings illustrate diverse scenarios and underscore economic, environmental, and social benefits, establishing our approach as an efficient decision-making tool across varied contexts.