Adaptive decision-making for civil infrastructure systems and communities exposed to evolving risks

Abstract Civil infrastructure systems and communities are continually subjected to changes in environmental and urban settings, evolving expectations and preferences of the public, tightening budgets, and unpredictable political circumstances over their service lives. Conventional decision methods, which determine optimal strategies based on stationary risk assessments and current knowledge, are not well suited for infrastructure systems/communities situated in evolving environments. There is a growing need to account for such evolutions in life-cycle risk assessment through adaptable, learning-based decision-making. Adaptive decision-making is a structured process of learning, improving understanding, and ultimately adapting management decisions in a systematic and efficient way, aimed at reducing uncertainties over the course of the management timeframe. This approach to risk management holds great potential for dealing with future challenges, by explicitly recognizing situational evolutions and improving decisions through learning. This paper investigates possible dynamic changes in the conditions of civil infrastructure systems/communities and their potential effects on life-cycle performance. A systematic adaptive decision process is proposed as a way to continually reevaluate the risks and provide more adaptive and flexible management actions to enhance infrastructure/community resilience under evolving conditions. Sequential Bayesian updating is utilized to reduce uncertainty and leverage the continuous accumulation of knowledge. Finally, the proposed adaptive decision methodology is illustrated with a benchmark problem based on a testbed residential community in Kathmandu, Nepal, to examine its feasibility and effectiveness in managing evolving risks. The results show that evolving exposure and vulnerability would increase future risks to the community and such risks could be effectively managed through adaptive decision-making.