A Multi-objective Genetic Algorithm for Inferring Inter-criteria Parameters for Water Supply Consensus

Abstract

Abstract. This work is based on a disaggregation approach for the ELECTRE III method for the group decision-making. We provide a procedure in which the group is supported for modifying the parameters of outranking methods in an iterative and interactive process. In this work, we provide an application of the procedure through evaluating eight municipal districts for Water Company to invest in projects of water supply. An inferring parameters model performed by NSGA-II obtains marginal information from decision makers with more disagreement, which supports the stage of parameters modification in correspondence with the preferences of the whole group (collective ranking). This paper shows how the inferring parameters model may be used to help the decision makers with different interest to iteratively reach an agreement on how to rank cities at a time, reflecting the preferences at the individual level and at the collective level.

An ELECTRE III based consensus‐reaching process to improve a collective solution

“…The supporting tool is based on a multiobjective genetic algorithm (MOGA) proposed by Alvarez et al. (2015) that infers intercriteria parameters, which are then proposed to each decision maker. For its input, this MOGA uses the individual and collective rankings as the preference information.…”

“…The GA borrows fundamental elements from NSGA II (Deb et al., 2002); see Alvarez et al. (2015) for a complete description of the MOGA for inferring intercriteria parameters. It optimizes the multiobjective combinatorial problem (6) to find sets of parameters that generate an individual ranking with more proximity to the group ranking.…”

An algorithm to elicitate ELECTRE II, III and IV parameters

“…Álvarez et al (2015) applied a multiobjective evolutionary algorithm, the NSGA-II (Non-dominated Sorting Genetic Algorithm II), to elicitate the criteria weights, indifference threshold, preference thresholds and veto thresholds of the ELECTRE III method. Their method included an interactive process of a group of experts that assisted them in finding a consensual set of parameters.…”

“…The dataset from Álvarez et al (2015) is going to be used to illustrate our methodology. This dataset, shown in Table 4, has eight actions and four criteria.…”

Robustness Analysis of an Outranking Model Parameters’ Elicitation Method in the Presence of Noisy Examples

“…[7,14]). EAs are powerful tools for the treatment of nonlinearity and global optimization in polynomial time [15], as a more recent example,Álvarez et al [16] used it to infer parameters that aid in the decision process at the collective level.…”