Evaluation of Flexible and Interactive Tradeoff Method Based on Numerical Simulation Experiments

Abstract

The Flexible and Interactive Tradeoff (FITradeoff) method is a multicriteria decision making/aiding (MCDM/A) method that uses partial information about the decision maker's (DM's) preferences in order to build a recommendation. This method preserves the strong axiomatic structure of the traditional tradeoff procedure, with an interactive and flexible process that demands less cognitive effort from the DM. Although FITradeoff has already been applied for aiding several practical decision situations, there is no previous study that tests the performance of this method with respect to expected theoretical benefits such as time and effort reduction. In this context, this paper presents the results of a simulation experiment that aims to analyze the behavior of FITradeoff in a wide range of scenarios. Computational constraints were identified in the simulation process, such as the number of simulations for each iteration. Despite the memory limitation of the software used, the number of total simulations performed was greater than what is commonly found in literature. We investigate how the performance of FITradeoff can be affected by changes in number of criteria, number of alternatives and weight pattern, and therefore it is possible to have a deeper understanding of the method and its main features. Therefore, this work focuses on studying, through simulations, the behavior of the multicriteria method FITradeoff, developed by De Almeida et al. (2016), filling a lack in the literature by analyzing said method in a wide array of scenarios by a simulation process in order to bring a better understanding of the method, as well as to validate it in a simulation study.

Using the FITradeoff method to solve a shopping mall location problem in the northeastern countryside of Brazil

“…Thus, it was used in medical applications (Dell'Ovo et al, 2020;Camilo et al 2020), in energy applications (Fossile et al, 2020;de Macedo, de Miranda-Mota and Sola, 2018;Kang, Frej and de Almeida, 2018), in environmental applications (Monte and Morais, 2019;Carrillo et al, 2018), in a security application (Camara e Silva et al, 2019), and in industry applications (Frej, de Almeida and Costa, 2017;Santos et al, 2020;Pergher et al, 2020; Silva, Costa and Frej, de Almeida and Costa, 2019;Lima, Viegas and Costa, 2017;de Gusmão and Pereira Medeiros, 2016). In order to test the performance of the FITradeoff method, Mendes et al (2020) performed simulation studies considering several scenarios, including different number of criteria and alternatives.…”

The use of the success‐based decision rule to support the holistic evaluation process in FITradeoff

“…It is worth mentioning that according to Mendes et al. (2020), in 81% of the cases, the subset of POAs is reduced to up to five alternatives after the ranking of criteria scaling constants. The LPP model used in the FITradeoff method is illustrated in the system of Equation (6): $$$$\begin{equation} \def\eqcellsep{&}\begin{array}{l} {\mathop{\rm Max}\nolimits}\displaystyle \sum_{i = 1}^n {{k_i}{v_i}} ({x_{ij}})\\[6pt] {\mathop{\rm s}\nolimits} .$$…”

“…Problems that had more than five alternatives have not been investigated in the neuroscience experiments, as according to Mendes et al. (2020), in 81% of the cases, the subset of POAs is reduced to up to five alternatives. Furthermore, holistic evaluation, such as those considered in this work, usually is not recommended for many alternatives due to the well‐known limitation of rationality in humans (Simon, 1995).…”

Multicriteria Model Based on FITradeoff Method for Prioritizing Sections of Brazilian Roads by Criticality

“…e procedure for eliciting criteria weights in the FITradeoff method is more flexible does not require direct information and requires less information from the DM. In fact, simulations performed by Mendes et al [43] show that, in 5% of the cases, a unique solution can be found after the criteria weights have been ranked using FITradeoff. Also, in 98% of the cases simulated, the subset of potentially optimal alternatives is reduced after the criteria weights have been ranked.…”