A genetic-algorithm-based method for tuning fuzzy logic controllers

doi:10.1016/S0165-0114(97)00309-6

Fuzzy Sets and Systems

Volume 108, Issue 1, 16 November 1999, Pages 39-47

https://doi.org/10.1016/S0165-0114(97)00309-6 Get rights and content

Abstract

It has been demonstrated many times in practice that fuzzy logic controllers have an important role in rule-based expert systems. However, it is essential for a fuzzy logic controller to have an appropriate set of rules to perform at the desired level. The linguistic structure of the fuzzy logic controller allows a tentative linguistic policy to be used as an initial rule base. At the design stage, if one can assemble a reasonably good collection of rules, it may then be possible to tune these rules to improve the controller performance. In this paper, a genetic-algorithm-based method for tuning the rule base of a fuzzy logic controller is presented. The method is used in tuning two PD-like fuzzy logic controllers and the results are discussed.

References (0)

Cited by (79)

Bee Colony Optimization metaheuristic for fuzzy membership functions tuning
2020, Expert Systems with Applications
The successfulness of the application of fuzzy logic to real-life problems depends on a number of parameters, such as the number and shapes of fuzzy membership functions, which are usually defined upon intuition or the subjective knowledge of relevant experts. One way to improve the performance of the fuzzy reasoning model is to use heuristic or metaheuristic techniques in order to optimize the shapes of membership functions. In this paper, the Bee Colony Optimization (BCO) algorithm is applied as a tool suitable for the specified optimization problem. The BCO belongs to the group of nature-inspired metaheuristics. The purpose of this paper is to present and discuss a strategy for the adjustment of fuzzy logic membership functions using the variant of the BCO algorithm based on the improvement of complete solutions and show its real-life application to the problem of the estimation of freight train energy consumption. According to the obtained results, it can be concluded that the precision of the developed fuzzy reasoning model is significantly increased after tuning membership functions by the BCO.
MULAN: Evaluation and ensemble statistical inference for functional connectivity
2018, NeuroImage
Citation Excerpt :
We used a genetic algorithm because it can handle the high degree of nonlinearity imposed by our strategy (traditional linear optimization tools cannot be used). Genetic algorithms have demonstrated the ability, efficiency and robustness in handling complex search spaces (Bancaud et al., 1970; Gürocak, 1999; Herrera et al., 1995) and achieving the complex objectives we set (the calculated connectivity graph needs to be similar to the ground-truth). However, other algorithms could be used and implemented.
Many analysis methods exist to extract graphs of functional connectivity from neuronal networks. Confidence in the results is limited because, (i) different methods give different results, (ii) parameter setting directly influences the final result, and (iii) systematic evaluation of the results is not always performed. Here, we introduce MULAN (MULtiple method ANalysis), which assumes an ensemble based approach combining multiple analysis methods and fuzzy logic to extract graphs with the most probable structure. In order to reduce the dependency on parameter settings, we determine the best set of parameters using a genetic algorithm on simulated datasets, whose temporal structure is similar to the experimental one. After a validation step, the selected set of parameters is used to analyze experimental data. The final step cross-validates experimental subsets of data and provides a direct estimate of the most likely graph and our confidence in the proposed connectivity. A systematic evaluation validates our strategy against empirical stereotactic electroencephalography (SEEG) and functional magnetic resonance imaging (fMRI) data.
Development of a Reinforcement Learning-based Evolutionary Fuzzy Rule-Based System for diabetes diagnosis
2017, Computers in Biology and Medicine
The early diagnosis of disease is critical to preventing the occurrence of severe complications. Diabetes is a serious health problem. A variety of methods have been developed for diagnosing diabetes. The majority of these methods have been developed in a black-box manner, which cannot be used to explain the inference and diagnosis procedure. Therefore, it is essential to develop methods with high accuracy and interpretability. In this study, a Reinforcement Learning-based Evolutionary Fuzzy Rule-Based System (RLEFRBS) is developed for diabetes diagnosis. The proposed model involves the building of a Rule Base (RB) and rule optimization. The initial RB is constructed using numerical data without initial rules; after learning the rules, redundant rules are eliminated based on the confidence measure. Next, redundant conditions in the antecedent parts are pruned to yield simpler rules with higher interpretability. Finally, an appropriate subset of the rules is selected using a Genetic Algorithm (GA), and the RB is constructed. Evolutionary tuning of the membership functions and weight adjusting using Reinforcement Learning (RL) are used to improve the performance of RLEFRBS. Moreover, to deal with uncovered instances, it makes use of an efficient rule stretching method. The performance of RLEFRBS was examined using two common datasets: Pima Indian Diabetes (PID) and BioSat Diabetes Dataset (BDD). The experimental results show that the proposed model provides a more compact, interpretable and accurate RB that can be considered to be a promising alternative for diagnosis of diabetes.
An improved genetic-fuzzy system for classification and data analysis
2017, Expert Systems with Applications
Citation Excerpt :
In the other approach, where interpretable Mamdani fuzzy systems are usually employed, some flexibility is added to the system’s structure to expand the search space and thus produce more accurate system (Casillas et al., 2003). Among the strategies applied to enhance the accuracy are: tuning the membership functions by changing the definition values of the parameters or their types (Casillas, Cordon, del Jesus, & Herrera, 2005; Cordon & Herrera, 1997; Gacto, Alcala, & Herrera, 2010; Gürocak, 1999; Jin, von Seelen, & Sendhoff, 1999; Nauck, 2000; Shi, Eberhart, & Chen, 1999), using linguistic modifiers in the rules that allow for more flexibility without losing the interpretability (Cordon, José del Jesus, & Herrera, 1998; Fernandez, del Jesus, & Herrera, 2010; Gonzalez & Perez, 1999; Herrera & Martinez, 2000) and learning the granularity of the fuzzy partitions to choose the level which gives more accuracy (Espinosa & Vandewalle, 2000; Gacto et al., 2010). Multi-objective evolutionary algorithms (MOEAs) have been extensively used in the context of interpretability to solve the problem of the interpretability-accuracy trade-off in fuzzy rule-based systems.
Interpretability of classification systems, which refers to the ability of these systems to express their behavior in an understandable way, has recently gained more attention and it is considered as an important requirement especially for knowledge-based systems. The main objective of this study is to improve the ability of a well-known fuzzy classifier proposed in Ishibuchi and Nojima (2007) to maximize the accuracy while preserve its interpretability. To achieve the above-mentioned objective, we propose two variants of the original fuzzy classifier. In the first variant classifier, the same components of the original classifier were used except NSGA-II which was replaced by an enhanced version called Controlled Elitism NSGA-II. This replacement aims at improving the ability of the first variant classifier to find non-dominated solutions with better interpretability-accuracy trade-off. In the second variant classifier, we further improve the first variant classifier by enhancing the selection method of the antecedent conditions of the rules generated in the initial population of genetic algorithm. Unlike the method applied in the original classifier and the first variant classifier, which uses a random selection of the antecedent conditions, we proposed a feature-based selection method to favor the antecedent conditions associated with the most relevant features. The results show that the two variant classifiers find more non-dominated fuzzy rule-based systems with better generalization ability than the original method which suggests that Controlled Elitism NSGA-II algorithm is more efficient than NSGA-II. In addition, feature-based selection method applied in the second variant classifier allowed this method to successfully obtain high-quality solutions as it has consistently achieved the best error rates for all the data sets compared to the original method and the first variant classifier.
Adaptive fuzzy multivariable controller design based on genetic algorithm for an air handling unit
2015, Energy
Citation Excerpt :
A method is presented in Ref. [17] for tuning fuzzy control rules by genetic algorithms to make the fuzzy logic control systems behave as closely as possible to the operator or expert behavior in a control process. An adaptive learning algorithm based on genetic algorithm for the tuning fuzzy rule base is presented in Ref. [18]. Two tuning algorithms namely Lateral and LA-based algorithms have been presented in Ref. [19] for the adaptation of fuzzy controllers.
In HVAC (Heating, Ventilation and Air Conditioning systems, effective thermal management is required because energy and operation costs of buildings are directly influenced by how well an air-conditioning system performs. HVAC systems are typically nonlinear time varying with disturbances, where conventional PID controllers may trade-off between stability and rise time. To overcome this limitation, a Genetic Algorithm based AFLC (Adaptive Fuzzy Logic Controller design has been proposed for the multivariable control of temperature and humidity of a typical AHU (air handling unit by manipulating valve positions to adjust the water and steam flow rates. Modulating equal percentage Globe valves for chilled water and steam have been modeled according to exact flow rates of water and steam. A novel method for the adaptation of FLC (Fuzzy Logic Controller by modifying FRM (Fuzzy Rule Matrix based on GA (genetic algorithm) has been proposed. This requires re-designing the complete FLC in MATLAB/Simulink whose procedure has also been proposed. The proposed adaptive controller outperforms the existing fuzzy controller in terms of steady state error, rise time and settling time.
Speed control of permanent magnet synchronous motors using fuzzy controller based on genetic algorithms
2012, International Journal of Electrical Power and Energy Systems
Citation Excerpt :
Overview of the studies on genetic fuzzy systems can be found in Cordon et al. [10] studies. Gürocak [12], in his study, proposed a method based on GA for generating rule base of a FLC. Shi and Eberhart [13] added types of membership functions into the used GA chromosome structure in addition to fuzzy membership function and rule base and hence, they formed a different chromosome structure.
In this study, speed control of a permanent magnet synchronous motor (PMSM) with genetic based fuzzy controller has been simulated. Earlier studies have focused on different components of fuzzy controllers such as rule base or data base in the literature. However, in this study, the whole knowledge base is parameterized and optimized to obtain an optimal fuzzy controller without expert knowledge. In the developed control scheme, there are three closed loops. The two inner loops are the current and the velocity feedbacks, respectively. The other outer feedback is the genetic algorithm (GA) which optimizes the rule base and data base of the fuzzy controller simultaneously. This outer loop is an iterative process. To make comparisons, a conventional fuzzy controller also has been designed and the PMSM speed control has been simulated for both the proposed and conventional fuzzy controller. The comparative results have proved that the proposed controller has a better dynamic response than that of the conventional one.

View all citing articles on Scopus

View full text