A multi-objective reliability-based decision support system for incorporating decision maker utilities in the design of infrastructure

摘要

Infrastructure comprises the most fundamental facilities and systems serving society. Because infrastructure exists in economic, social, and environmental contexts, all lifecycle phases of such facilities should maximize utility for society, occupants, and designers. However, due to uncertainties associated with the nature of the built environment, the economic, social, and environmental (i.e., triple bottom line) impacts of infrastructure assets must be described as probabilistic. For this reason, optimization models should aim to maximize decision maker utilities with respect to multiple and potentially conflicting probabilistic decision criteria. Although stochastic optimization and multi-objective optimization are well developed in the field of operations research, their intersection (multi-objective optimization under uncertainty) is much less developed and computationally expensive. This article presents a computationally efficient, adaptable, multi-objective decision support system for finding optimal infrastructure design configurations with respect to multiple probabilistic decision criteria and decision maker requirements (utilities). The proposed model utilizes the First Order Reliability Method (FORM) in a systems reliability approach to assess the reliability of alternative infrastructure design configurations with regard to the probabilistic decision criteria and decision maker defined utilities, and prioritizes the decision criteria that require improvement. A pilot implementation is undertaken on a nine-story office building in Los Angeles, California to illustrate the capabilities of the framework. The results of the pilot implementation revealed that "high-performing" design configurations (with higher initial costs and lower failure costs) had a higher probability of meeting the decision maker's preferences than more traditional, low initial cost configurations. The proposed framework can identify low-impact designs that also maximize decision maker utilities.