Optimization-based multicriteria decision analysis for identification of desired petroleum-contaminated groundwater remediation strategies.

Abstract

The conventional multicriteria decision analysis (MCDA) methods used for pollution control generally depend on the data currently available. This could limit their real-world applications, especially where the input data (e.g., the most cost-effective remediation cost and eventual contaminant concentration) might vary by scenario. This study proposes an optimization-based MCDA (OMCDA) framework to address such a challenge. It is capable of (1) capturing various preferences of decision-makers, (2) screening and analyzing the performance of various optimized remediation strategies under changeable scenarios, and (3) compromising incongruous decision analysis results. A real-world case study is employed for demonstration, where four scenarios are considered with each one corresponding to a set of weights representative of the preference of the decision-makers. Four criteria are selected, i.e., optimal total pumping rate, remediation cost, contaminant concentration, and fitting error. Their values are determined through running optimization and optimization-based simulation procedures. Four sets of the most desired groundwater remediation strategies are identified, implying specific pumping rates under varied scenarios. Results indicate that the best action lies in groups 32 and 16 for the 5-year, groups 49 and 36 for the 10-year, groups 26 and 13 for the 15-year, and groups 47 and 13 for the 20-year remediation.