This section depicts, the literature review of published paper from the domain of applications of MANFIS technique in various fields and fault diagnosis.
A neuro-fuzzy inference system, or equivalently, a neuro-fuzzy system is a fuzzy inference system which employs neural network learning techniques. Multiple adaptive neuro-fuzzy inference system (MANFIS) [l27, 128, 129] is an extension of a single-output neuro-fuzzy system, ANFIS, so that multiple outputs can be handled. A neuro-fuzzy system is a nonparametric regression tool, which models the regression relationship non-parametrically without reference to any pre-specified functional form, and it is capable of modeling highly nonlinear and approximately known systems.
Cheng et al. [130] have optimize a multiple output system using the MANFIS neuro-fuzzy network for modeling the system and genetic algorithm has been used to optimize the multiple objective function. The validity of the technique has been performed using a practical problem. Buyukozkan et al. [131] have studied the performance of a new product development process (NPD) under uncertain conditions and given their effort to improve the quality of decision-making in NPD by following new iterative methodology. They have used fuzzy logic, neural networks and MANFIS technique for improvising the methodology for new product idea selection. Hengjie et al. [132] have presented a probabilistic fuzzy neural network (ProFNN) approach for handling randomness in the system by introducing the probability of input linguistic terms and providing linguistic meaning into the connectionist
architecture. The results from the proposed technique have been compared with that of multi-input–multi-output-ANFIS (MANFIS), self-organizing adaptive fuzzy neural control and Extreme Learning Machine for validation of the probabilistic fuzzy neural network. Vairappan et al. [133] have illustrated an improved adaptive neuro-fuzzy inference system (ANFIS) for the application of time-series prediction. The proposed improved version of ANFIS has introduced the application of self-feedback connections for modeling the temporal dependence. The effectiveness of the proposed methodology has been validated by using three benchmark time-series tests. Gholamian et al. [134] have presented a systematic design for multi objective problems using hybrid intelligent system to solve ill-structured situations. Fuzzy rules and neural networks are used in this systematic design and the developed hybrid system is established with the ability of mapping between objective space and solution space. The results obtained are authenticated on three test problems. Ellithy et al. [135] presented a methodology based on ANFIS to improve the damping of power systems in the presence of load model parameters uncertainty. The proposed ANFIS is trained over a wide range of typical load parameters to adapt the gains of the SVC stabilizer. They have claimed that the simulation results are showing encouraging trends in comparison to SVC stabilizer operating on other techniques. Güneri et al. [136] have developed a new approach to address the supplier selection problem. The proposed ANFIS model has been trained with parameters relating to supplier selection criteria. They have tested the results from their technique by comparing with the results of the multiple regression method, demonstrating that the ANFIS method performed well. Nagarajan et al. [137] in their study have proposed the design of Adaptive Neuro-Fuzzy Observer based sensor fault detection in a three-tank interacting level process. They have designed the fault detection algorithm with Multiple Adaptive Neuro-Fuzzy Inference System (MANFIS) that uses a neural network to fix optimal shape and parameters for the membership functions and effective rule base for the fuzzy system. Fault detection is being performed by them estimating the states of the level process and comparing them with measured values. Jassar et al. [138] have established a technique to find out the temperature in heating space utilizing an adaptive neuro-fuzzy inference system. The proposed system has been developed by combining the fuzzy inference systems and artificial neural networks. The results from the developed method have been cross verified by experimentation. Asensi et al. [139] have formulated a system
based on multiple adaptive neuro-fuzzy inference systems (MANFIS) to analyze the performance characteristics of analog circuit. Zhang et al. [140] have studied a dynamic system and developed an algorithm to identify the chaotic signals present in a system by adopting adaptive-neuro-fuzzy-inference system (ANFIS) and MANFIS methodology. Nguyen et al. [141] have used vibration analysis and fuzzy logic technique to develop a fault detection method in bearings. The parameters representing the condition of the system have been used to design the proposed technique based on Adaptive Network based Fuzzy Inference System (ANFIS) and Genetic Algorithm (GA). The results obtained from the developed model have been tested with other set of bearing data to exhibit the reliability of the chosen model. Lei et al. [142,143] have proposed a method for fault diagnosis of rolling element bearing system using multiple adaptive neuro-fuzzy inference systems (MANFIS) and empirical mode decomposition (EMD). The robustness of the developed mechanism has been checked by employing the same on different bearing systems.
So in the subsequent section algorithm have been discussed used for fault diagnosis using hybrid AI techniques such as Neuro-Fuzzy, Genetic-fuzzy Technique, Genetic-neural Technique and Genetic-neural-fuzzy Technique.