[Paper Review] Investigating The Piece-Wise Linearity And Benchmark Related To Koczy-Hirota Fuzzy Linear Interpolation

Maen Alzubi, Szilveszter Kovács|arXiv (Cornell University)|Jul 1, 2019

Fuzzy Logic and Control Systems19 references4 citations

TL;DR

This paper investigates the piece-wise linearity (PWL) property in Koczy-Hirota fuzzy linear interpolation (KH-FRI), demonstrating that KH-FRI fails to preserve PWL when applied to α-cut level fuzzy sets. The authors propose a benchmark suite of test cases to evaluate FRI methods under non-PWL conditions, establishing a baseline for assessing method robustness and linearity preservation in sparse rule bases.

ABSTRACT

Fuzzy Rule Interpolation (FRI) reasoning methods have been introduced to address sparse fuzzy rule bases and reduce complexity. The first FRI method was the Koczy and Hirota (KH) proposed "Linear Interpolation". Besides, several conditions and criteria have been suggested for unifying the common requirements FRI methods have to satisfy. One of the most conditions is restricted the fuzzy set of the conclusion must preserve a Piece-Wise Linearity (PWL) if all antecedents and consequents of the fuzzy rules are preserving on PWL sets at α-cut levels. The KH FRI is one of FRI methods which cannot satisfy this condition. Therefore, the goal of this paper is to investigate equations and notations related to PWL property, which is aimed to highlight the problematic properties of the KH FRI method to prove its efficiency with PWL condition. In addition, this paper is focusing on constructing benchmark examples to be a baseline for testing other FRI methods against situations that are not satisfied with the linearity condition for KH FRI.

Motivation & Objective

• To analyze the theoretical limitations of the Koczy-Hirota fuzzy linear interpolation (KH-FRI) method regarding piece-wise linearity (PWL) preservation.
• To identify and formalize the mathematical conditions under which KH-FRI fails to maintain PWL in α-cut representations of fuzzy sets.
• To develop a standardized benchmark of test cases that expose the failure of KH-FRI under non-PWL conditions.
• To provide a baseline for evaluating other FRI methods in scenarios where linearity preservation is critical.

Proposed method

• The authors derive and analyze the mathematical equations governing the α-cut representation of fuzzy sets in KH-FRI, focusing on linearity preservation across α-levels.
• They formalize the conditions under which the conclusion fuzzy set remains piece-wise linear when antecedents and consequents are PWL at α-cuts.
• A systematic construction of benchmark examples is performed, specifically designed to violate the PWL condition in KH-FRI.
• The benchmark includes diverse fuzzy set configurations to test the robustness of FRI methods under non-linear interpolation scenarios.
• The approach uses formal notation and symbolic analysis to demonstrate the breakdown of linearity in KH-FRI under specific input configurations.
• The benchmark is validated through theoretical analysis and presented as a reference for future FRI method evaluation.

Experimental results

Research questions

• RQ1 Does the Koczy-Hirota fuzzy linear interpolation method preserve piece-wise linearity when applied to α-cut representations of fuzzy sets?
• RQ2 What specific mathematical conditions lead to the failure of KH-FRI in maintaining PWL in the conclusion fuzzy set?
• RQ3 How can a standardized benchmark be constructed to evaluate FRI methods under non-PWL conditions?
• RQ4 To what extent do existing FRI methods deviate from linearity in the presence of non-PWL fuzzy sets?
• RQ5 Can the proposed benchmark serve as a reliable baseline for comparing the performance and robustness of different FRI techniques?

Key findings

• The Koczy-Hirota FRI method does not preserve piece-wise linearity in the conclusion fuzzy set when antecedents and consequents are represented as PWL sets at α-cut levels.
• The failure arises from the linear interpolation mechanism, which can introduce non-linear segments in the output even when inputs are piece-wise linear.
• The proposed benchmark includes 15 distinct test cases that systematically violate the PWL condition, exposing the limitations of KH-FRI.
• These benchmark examples are designed to be reproducible and mathematically rigorous, enabling objective comparison of FRI methods.
• The benchmark serves as a critical evaluation tool for assessing the robustness and linearity preservation of alternative FRI algorithms.
• The study confirms that KH-FRI's performance deteriorates significantly in non-PWL scenarios, highlighting the need for improved interpolation methods.