Reinforcement learning for decision-making under deep uncertainty.

Planning under complex uncertainty often asks for plans that can adapt to changing future conditions. To inform plan development during this process, exploration methods have been used to explore the performance of candidate policies given uncertainties. Nevertheless, these methods hardly enable adaptation by themselves, so extra efforts are required to develop the final adaptive plans, hence compromising the overall decision-making efficiency. This paper introduces Reinforcement Learning (RL) that employs closed-loop control as a new exploration method that enables automated adaptive policy-making for planning under uncertainty. To investigate its performance, we compare RL with a widely-used exploration method, Multi-Objective Evolutionary Algorithm (MOEA), in two hypothetical problems via computational experiments. Our results indicate the complementarity of the two methods. RL makes better use of its exploration history, hence always providing higher efficiency and providing better policy robustness in the presence of parameter uncertainty. MOEA quantifies objective uncertainty in a more intuitive way, hence providing better robustness to objective uncertainty. These findings will help researchers choose appropriate methods in different applications.

Substantially reducing global PM2.5-related deaths under SDG3.9 requires better air pollution control and healthcare.

The United Nations' Sustainable Development Goal (SDG) 3.9 calls for a substantial reduction in deaths attributable to PM2.5 pollution (DAPP). However, DAPP projections vary greatly and the likelihood of meeting SDG3.9 depends on complex interactions among environmental, socio-economic, and healthcare parameters. We project potential future trends in global DAPP considering the joint effects of each driver (PM2.5 concentration, death rate of diseases, population size, and age structure) and assess the likelihood of achieving SDG3.9 under the Shared Socioeconomic Pathways (SSPs) as quantified by the Scenario Model Intercomparison Project (ScenarioMIP) framework with simulated PM2.5 concentrations from 11 models. We find that a substantial reduction in DAPP would not be achieved under all but the most optimistic scenario settings. Even the development aligned with the Sustainability scenario (SSP1-2.6), in which DAPP was reduced by 19%, still falls just short of achieving a substantial (≥20%) reduction by 2030. Meeting SDG3.9 calls for additional efforts in air pollution control and healthcare to more aggressively reduce DAPP.

Co-creating local socioeconomic pathways for achieving the sustainable development goals.

The Sustainable Development Goals (SDGs) recognise the importance of action across all scales to achieve a sustainable future. To contribute to overall national- and global-scale SDG achievement, local communities need to focus on a locally-relevant subset of goals and understand potential future pathways for key drivers which influence local sustainability. We developed a participatory method to co-create local socioeconomic pathways by downscaling the SDGs and driving forces of the shared socioeconomic pathways (SSPs) via a local case study in southern Australia through contextual analysis and community engagement. We linked the SSPs and SDGs by identifying driving forces and describing how they affect the achievement of local SDGs. We co-created six local socioeconomic pathways with the local community which track towards futures with different levels of fulfilment of the SDGs and each encompasses a narrative storyline incorporating locally-specific ideas from the community. We tested and validated the local pathways with the community. This method extends the SSPs in two dimensions-into the broader field of sustainability via the SDGs, and by recontextualizing them at the local scale. The local socioeconomic pathways can contribute to achieving local sustainability goals from the bottom up in alignment with global initiatives.