Real-Time Synergies between Homeostatic Technological and Homeorhetic Ecological Systems by Multiscale MPC and Bayesian Optimization

Abstract

Engineering design and operations have traditionally favored steady-state optimization, often overlooking the dynamic and intermittent nature of ecosystems. This has led to environmental degradation and unsustainable practices. Techno-Ecological Synergy (TES) offers an alternative framework that seeks to harmonize technological systems with ecological processes, but previous TES studies have relied on retrospective models that assume perfect foresight of ecosystem behavior. This work presents a novel framework called TES-IDC (Techno-Ecological Synergy - integrated design and control) that addresses the limitations of retrospective TES models by incorporating adaptive recourse decisions. The framework extends the IDC methodology to TES models, utilizing a simulation-optimization approach to separate design and operational problems. The operational problem is modeled as a closed-loop model predictive controller (MPC) simulation, while Bayesian optimization is employed to identify design conditions that minimize both capital and operational costs. A key innovation of TES-IDC is the use of an infinite-horizon MPC policy to account for both short- and long-term impacts at a reasonable computational cost. The effectiveness of the TES-IDC framework is demonstrated through an air quality regulation case study, where it determines the optimal size of a reforestation area and a policy for technological operations, considering the environment’s dynamic capacity. The derived operational policy successfully meets short-term air quality constraints while optimizing long-term economic and ecological objectives. This research highlights the potential of TES-IDC in designing sustainable systems that adapt to the dynamic nature of ecosystems, paving the way for a more harmonious coexistence between human activities and the environment.