Standard Interactive Genetic Algorithm—Comprehensive Optimization Framework for Groundwater Monitoring Design

Abstract

Optimization for water resources management typically requires many simplifying assumptions about the definition and characteristics of the policy or design application in order to express decision makers’ criteria as mathematical objectives and constraints. However, real-world applications often involve important subjective information that cannot be reflected in mathematical expressions accurately or completely. This can result in mathematically optimized solutions that are less meaningful or desirable to decision makers. To address this issue, this paper presents the standard interactive genetic algorithm (SIGA) methodology that enables human decision makers to effectively analyze subjective information that is not easily quantifiable and make decisions about the quality of a design based on their preferences. These decisions are used as continuous run-time subjective feedback, along with the mathematically defined objectives and constraints, to search for optimal designs that reflect both quantitative and qualitative objectives. Although this interactive optimization methodology is applicable for any water resources planning and management problems, this paper focuses on exploring the benefits of such an approach within the domain of groundwater monitoring design. Systematic procedures and guidelines for designing a SIGA are presented, along with proposed strategies for improving the performance of SIGA. The SIGA approach is also compared with a noninteractive genetic algorithm strategy for a real-world application, and the advantages and limitations of the interactive strategy are examined.