Against Artificial Complexification: Crisp vs. Fuzzy Information in the TOPSIS Method

Abstract

The question of whether the use of crisp or fuzzy input information in the TOPSIS method produces a different ranking is explored. Using a basic representation of fuzziness through triangular fuzzy numbers, a set of randomly generated fuzzy and crisp multicriteria decision problems are solved. Then, the corresponding rankings are compared and variations in the top alternative are studied.The results show that changes in the top alternative are minor. This situation, coupled with the fact that the "true" ranking is unknown and that more complex models of "fuzzy" information require a huge amount of precise information from the decision maker side, raise the discussion of whether in this specific context a "complexification" of the input data makes sense.

Forecasting Pergerakan Harga Volatility Index dengan Menggunakan Metode Fuzzy Tsukamoto dan Evaluasi Dstat Metric

“…Secara konsep teori, didalam kecerdasan buatan terdapat logika crisp dan fuzzy yang dapat digunakan untuk mengolah data historical [3]. Hal ini juga mendasari beberapa penelitian untuk melakukan forecasting didalam dunia finance.…”

“…Untuk nilai periode yang digunakan dalam indikator Relative Strength Index sebesar 14 [14] [17]. Sedangkan untuk periode Stochastic Oscillators adalah default (14,3,3) [15]. Beberapa penelitian tersebut memiliki perbedaan dengan penelitian ini, yaitu penggunaan logika fuzzy yang akan digunakan.…”