Evolving Compact and Interpretable Takagi–Sugeno Fuzzy Models With a New Encoding Scheme

Abstract

Developing Takagi-Sugeno fuzzy models by evolutionary algorithms mainly requires three factors: an encoding scheme, an evaluation method, and appropriate evolutionary operations. At the same time, these three factors should be designed so that they can consider three important aspects of fuzzy modeling: modeling accuracy, compactness, and interpretability. This paper proposes a new evolutionary algorithm that fulfills such requirements and solves fuzzy modeling problems. Two major ideas proposed in this paper lie in a new encoding scheme and a new fitness function, respectively. The proposed encoding scheme consists of three chromosomes, one of which uses unique chained possibilistic representation of rule structure. The proposed encoding scheme can achieve simultaneous optimization of parameters of antecedent membership functions and rule structures with the new fitness function developed in this paper. The proposed fitness function consists of five functions that consider three evaluation criteria in fuzzy modeling problems. The proposed fitness function guides evolutionary search direction so that the proposed algorithm can find more accurate compact fuzzy models with interpretable antecedent membership functions. Several evolutionary operators that are appropriate for the proposed encoding scheme are carefully designed. Simulation results on three modeling problems show that the proposed encoding scheme and the proposed fitness functions are effective in finding accurate, compact, and interpretable Takagi-Sugeno fuzzy models. From the simulation results, it is shown that the proposed algorithm can successfully find fuzzy models that approximate the given unknown function accurately with a compact number of fuzzy rules and membership functions. At the same time, the fuzzy models use interpretable antecedent membership functions, which are helpful in understanding the underlying behavior of the obtained fuzzy models.

A new approach to nonlinear modelling of dynamic systems based on fuzzy rules

“…In literature we can find many methods to design a structure and select parameters of neuro-fuzzy systems (see, e.g., Kim et al, 2006;Angelov and Filev, 2004;Medasani et al, 1998;Rutkowski and Cpałka, 2005;Starczewski et al, 2010;Malchiodi and Pedrycz, 2013;Cpałka, 2009a;2009b;Cpałka et al, 2014;2013). In this paper we used the (λ + μ) evolutionary strategy, which belongs to the group of population based algorithms.…”

Developing a bioaerosol detector using hybrid genetic fuzzy systems

“…Instead of those aforementioned metrics, slightly modified versions of interpretability metrics proposed in Kim et al (2006), namely the length of overlap and the length of discontinuity between fuzzy sets, are used. In a nutshell, it is desired that the intersection value of two fuzzy sets would lie between user specified constants a L and a H .…”

“…First, calculating P OL and P DC is briefly presented, and reader is referred to Kim et al (2006) for more information. Finally, computing P MV is introduced.…”

“…To calculate the length of overlap for fuzzy sets, left and right intersection points for the user specified level a H ¼ 0:6 (the same value as in Kim et al, 2006) are computed first for all fuzzy sets. Since gbell MFs are applied, the intersection points can be computed as …”

Automatically extracting T–S fuzzy models using cooperative random learning particle swarm optimization

“…The fuzzy model design can be considered as an optimization problem in multidimensional space where each point represents a potential fuzzy model with different structures and the related parameters [21,30]. Because of their powerful global searching capability, evolutionary algorithms (EAs), such as genetic algorithm [5], genetic programming [6], evolutionary programming [16], evolution strategy [27] and differential evolution [9], have been used to construct fuzzy models.…”