A grid-based optimization algorithm for parameters elicitation in WOWA operators: An application to risk assesment

In this paper we propose a joint grid-based and stochastic search for parameters elicitation in the case of WOWA aggregation functions. The method uses a grid search approach to determine the parameter of a monotonic quantifier, and for each of the values, a stochastic search in the space of the criteria weights minimizes the sum of the quadratic error between the computed and the real output of a learning set. A simulated application is proposed in the case of transportation risk assessment, using an ad hoc questionnaire applied to risk matrices.