Human-in-the-loop MGA to generate energy system design options matching stakeholder needs

Francesco Lombardi,Stefan Pfenninger

2024-07-19

Abstract:Using cost-optimisation to support energy system design decisions hides from view many more feasible designs that, while economically comparable to the least-cost option, may have other practical advantages that stakeholders prefer. Increasingly, modellers look at Modelling to generate alternatives (MGA) as a way to go beyond few, optimal designs and provide stakeholders with broad portfolios of design options to appraise based on their preferences and knowledge of the problem. In practice, however, generating and exploring all the feasible system design options is not computationally possible. Modellers need to choose what design feature to generate diversity around, and yet, they cannot know what features may matter the most to heterogeneous real-world stakeholders. Stakeholders' knowledge is thus needed to guide the MGA search towards the most important trade-offs, but how to incorporate such knowledge in the MGA workflow is unclear. To address this gap, we propose a first-of-its-kind human-in-the-loop (HITL) approach that seamlessly and automatically integrates stakeholder preferences into an MGA workflow. In a nutshell, we elicit stakeholders' system design preferences via a first interaction between them and a tentative MGA design space. Hence, we decode those preferences so that it becomes possible to feed them back to the MGA algorithm to perform a guided search. This search produces an updated, human-trained design space with a higher share of designs that align with the elicited preferences. In a controlled experiment for the case study of the Portuguese energy system, we demonstrate that our HITL-MGA approach may facilitate the identification of a consensus solution that balances many conflicting stakeholder preferences. The approach shows promise for generalisation and real-world application to accelerate technically and socially feasible energy transition decisions.

Physics and Society

What problem does this paper attempt to address?

This paper aims to address key challenges in energy system design decision-making, particularly in the generation of energy system design solutions that balance technical feasibility and social acceptance when considering the preferences and knowledge of multiple stakeholders. Traditionally, cost optimization methods have been used to support energy system design decisions, but they often overlook many other viable designs that may be economically competitive with the lowest-cost options but more feasible in practice due to other unmodeled benefits. To go beyond a limited number of optimal solutions, researchers have turned to the "modeling to generate alternatives" (MGA) approach, which can provide a wide range of design options for stakeholders to assess based on their preferences and understanding of the problem.

However, due to limited computational capacity, it is practically impossible to generate and explore all feasible system design options. Therefore, it is necessary to select which design features to focus on to generate diversity. Additionally, researchers cannot predict which features are most important in the real world for stakeholders with different needs. Thus, stakeholder knowledge is crucial in guiding the MGA search towards the most significant trade-offs, but how to integrate this knowledge into the MGA workflow remains unclear.

To bridge this gap, this paper proposes an innovative human-in-the-loop (HITL) approach that seamlessly and automatically integrates stakeholder preferences into the MGA workflow. Firstly, stakeholder preferences for system design are collected through their initial interaction with the MGA design space. Then, these preferences are decoded to provide feedback to the MGA algorithm, guiding the search to generate an updated, human-informed design space that contains more designs matching the collected preferences.

Through a controlled experiment in the case study of the Portuguese energy system, the paper demonstrates that this HITL-MGA approach can facilitate the identification of a consensus solution that reconciles conflicting stakeholder preferences. This method demonstrates potential in terms of generality and practical application, and it is expected to accelerate technically and socially viable energy transition decisions. In conclusion, the objective of the paper is to develop a HITL-MGA approach that effectively utilizes stakeholder knowledge to tailor the search strategy of the MGA algorithm, thereby improving the quality of the design space and the efficiency of the decision-making process.