4 - Soft computing techniques

The past few years have witnessed a growing rate of attraction in adoption of Artificial Intelligence (AI) techniques to solve different engineering problems. Besides, Short Term Electrical Load Forecasting (STLF) is one of the important concerns of power systems and accurate load forecasting is vital for managing supply and demand of electricity. This study estimates short term electricity loads of Iran by means of Adaptive Neuro-Fuzzy Inference System (ANFIS), Artificial Neural Networks (ANN) and Genetic Algorithm (GA) which are the most successful AI techniques in this field. In order to improve forecasting accuracy, all AI techniques are equipped with preprocessing concept, and effects of this concept on performance of each AI technique are investigated. Finally, outcomes of the approaches are evaluated and compared by means of the mean absolute percentage error (MAPE). Results show that data preprocessing can significantly improve performance of the AI techniques. Meanwhile, ANFIS outcomes are more approximate to the actual loads than those of ANN and GA, so it can be considered as a suitable tool to deal with STLF problems.

In this study, we examined the ability of Artificial Intelligence (AI) techniques in deriving parameters of NMR log starting from conventional logs. To perform this, it was applied Fuzzy Logic (FL) and Artificial Neural Networks (ANN) techniques separately, forming independent schemes. On the other hand, Genetic Algorithms (GA) and Simple Mean (SM) approaches were used to assign weighting factors to FL and ANN estimates, with the objective to optimize the individual contributions of each one. This methodology made use of, as input data, gamma rays (GR), resistivity laterolog (RLA1), density (RHOB), neutron (NPOR) and sonic (DTCO) logs from two wells drilled through a carbonate reservoirs in Campos Basin - Brazil. The output responses were compared with free fluid porosity (CMFF) and SDR lateral permeability (KSDR - Schlumberger Doll Research), both derived from NMR log of the same wells. The results indicate that ANN performed better when compared with FL, but this last was essential in the success of SM and GA estimates. However, each approach showed a good fit with the parameter curves of NMR log, confirming the utility of the present methodology in the case when there are only conventional logs in the studied well.

A comprehensive overview on signal processing and artificial intelligence techniques applications in classification of power quality disturbances

Advances in artificial intelligence (AI) techniques that can be used for different purposes have enabled it to be used in many different industrial applications. These are mainly used for modeling, identification, optimization, prediction and control of complex systems under the influence of more than one parameter in industrial applications. With the increasing accuracy of AI techniques, it has also obtained a wide application area on friction stir welding (FSW), one of the production methods developed in recent years. In this study, commonly used AI techniques for FSW, results, accuracy and superiority of AI techniques are reviewed and evaluated. In addition, an overview of AI techniques for FSW in different material combinations is provided. Considering the articles examined; It is seen that welding speed, rotational speed, the plunge depth, spindle torque, shoulder design, base material, pin design/profile, tool type are used as input parameters and tensile strength, yield strength, elongation, hardness, wear rate, welding quality, residual stress, fatigue strength are used as output parameters. As can be seen from the studies, it made important contributions in deciding what input parameters should be in order to have the output parameter at the desired value. The most common used materials for FSW are Al, Ti, Mg, Brass, Cu and so on. When FSW studies using artificial intelligence techniques were examined, it was seen that 81% of the most used materials were AL alloys and 23% of them were made with dissimilar materials. The most commonly utilized AI techniques were said to be artificial neural networks (ANN), fuzzy logic, machine learning, meta-heuristic methods and hybrid systems. As a result of the examination, ANN was the most widely used method among these methods. However, in recent years, with the exploration of new hybrid methods it was seen that hybrid systems used with ANN have higher accuracy.

Recognition of the importance of using artificial neural networks and genetic algorithms to optimize chiller operation

Development of new correlations for the oil formation volume factor in oil reservoirs using artificial intelligent white box technique

Exploring the effect of basalt fibers on maximum deviator stress and failure deformation of silty soils using ANN, SVM and FL supported by experimental data

This study presents a comprehensive analysis of artificial intelligence (AI) techniques to predict healthcare expenditure per capita (pcHCE) in Turkey. Well‐known AI techniques such as random forest (RF), artificial neural network (ANN), multiple linear regression (MLR), support vector regression (SVR), and relevance vector machine (RVM) were used to forecast pcHCE. Twenty‐nine years of historical data from 1990 to 2018 were used in the training and testing phases of the models. Gross domestic product per capita, life expectancy at birth, unemployment rate, crude birth rate, and the number of physicians and hospitals were used as input variables for the analysis. A genetic algorithm‐based feature selection (GAFS) method was applied to all models to select the relevant and optimal feature subset in the prediction of pcHCE. The comparative results showed that the GAFS method improved the overall performance of all base AI models. The hybrid GAFS‐RF model performed best among all AI‐based prediction methods, with a 99.86% correlation of determination (R2) value at the testing stage.

Shear connectors play a prominent role in the design of steel-concrete composite systems. The behavior of shear connectors is generally determined through conducting push-out tests. However, these tests are costly and require plenty of time. As an alternative approach, soft computing (SC) can be used to eliminate the need for conducting push-out tests. This study aims to investigate the application of artificial intelligence (AI) techniques, as sub-branches of SC methods, in the behavior prediction of an innovative type of C-shaped shear connectors, called Tilted Angle Connectors. For this purpose, several push-out tests are conducted on these connectors and the required data for the AI models are collected. Then, an adaptive neuro-fuzzy inference system (ANFIS) is developed to identify the most influencing parameters on the shear strength of the tilted angle connectors. Totally, six different models are created based on the ANFIS results. Finally, AI techniques such as an artificial neural network (ANN), an extreme learning machine (ELM), and another ANFIS are employed to predict the shear strength of the connectors in each of the six models. The results of the paper show that slip is the most influential factor in the shear strength of tilted connectors and after that, the inclination angle is the most effective one. Moreover, it is deducted that considering only four parameters in the predictive models is enough to have a very accurate prediction. It is also demonstrated that ELM needs less time and it can reach slightly better performance indices than those of ANN and ANFIS.

Artificial Intelligence-based models are widely used in health care for prediction, detection and treatment of diseases. The aim of this paper is to study artificial intelligence (AI) techniques used in ophthalmology. Many AI-based models are used to solve ophthalmological issues and show good results for detection of eye diseases like glaucoma and diabetic retinopathy, age-related macular degeneration, amblyopia, etc. Amblyopia is an eye disorder in which brain catches the signals from one eye only and starts ignoring the signals from the other eye, which is called as a lazy eye. Lazy eye should be detected on time for its timely treatment, otherwise its treatment is not possible in adulthood. Earlier amblyopia was detected with traditional method, but now early detection is possible with different machine and deep learning techniques. In this paper, we did exhaustive study of artificial intelligence techniques such as decision tree, random forest and artificial neural network, k-nearest neighbor, Naive Bayes used by many researchers for the detection of amblyopia. Pros and cons of each method are also discussed. From this survey, it is concluded that there has not been much work done for the detection of amblyopia (lazy eye), and it needs further research to get more accuracy in the detection models.

Chapter 16 - Modeling the Axial Capacity of Bored Piles Using Multi-Objective Feature Selection, Functional Network and Multivariate Adaptive Regression Spline

Composite materials offer a wide range of advantages in terms of strength, weight, and versatility, making them a popular choice for structural applications. The performance of composite structures depends not only on the material constituents but also on the topology and arrangement of these components. However, designing an optimal configuration that maximizes the desired performance metrics can be a challenging task due to the large design space and complex interactions between the components. It presents a novel approach for optimizing the multi-component topology and material orientation design of composite structures using artificial intelligence (AI) techniques. The objective is to develop an automated and efficient methodology that can identify the most optimal configuration for a given set of performance requirements. The proposed methodology combines AI algorithms, such as genetic algorithms and machine learning, with advanced computational modeling techniques. The genetic algorithms are employed to explore the design space and search for the best combination of topology and material orientations. Concurrently, machine learning techniques are utilized to model the complex relationships between the design variables and performance metrics, enabling the identification of design patterns and accelerating the optimization process.
 To achieve this, a comprehensive framework is established, encompassing the generation of an initial population of candidate designs, evaluation of their performance using numerical simulations or experimental testing, application of genetic algorithms to iteratively evolve the population by selecting the fittest designs and introducing variations, and utilization of machine learning models to predict the performance of new designs and guide the optimization process. The effectiveness of the proposed methodology is demonstrated through a case study involving the design of a composite aerospace structure. The results reveal that the AI-based optimization approach significantly outperforms traditional trial-and-error methods, leading to improved performance metrics such as strength, stiffness, and weight.
 The findings have significant implications for the design and manufacturing of composite structures in various industries, including aerospace, automotive, and civil engineering. The automated optimization process enabled by AI techniques can efficiently explore the design space, identify innovative solutions, and ultimately enhance the overall performance of composite structures.

In the design process, it would be much better to give designers more and more acceptable and reasonable design candidates to derive their preferences to meet requirements of multiple high performances. In this paper, we will propose a new method, to give such acceptable design candidates based on genetic algorithms (GAs) with considering something like strategy for adaptation. GAs are search algorithms based on the mechanics of natural selection and natural genetics. Yet, they are no simple random walk but they efficiently exploit historical information to speculate on new search points with expected improved performance. Thus, we expect in GAs to give multiple acceptable and near optimum design candidates just like so many species in natural living things. However, in simple GAs, after specific number of generations, their populations become near one or two specific local optimum solutions. (Hoping for including global optimum solution.) In order to obtain multiple acceptable solutions, we need to perform GAs processes by keeping variations in characters of individuals. The proposed method tries to find multiple acceptable solutions by considering strategy for adaptation into GAs processes, which are food chain, strategy of foraging, death strategy and strategy of reproduction. As a numerical example, we apply the proposed method to simple multi-objective optimization and demonstrate its efficiencies.

An artificial neural network (ANN) is an information processing modelling of the human brain inspired by the way biological nervous systems behave. There are about 100 billion neurons in the human brain. Each neuron has a connection point between 1,000 and 100,000. The key element of this paradigm is the novel structure of the information processing system. In the human brain, information is stored in such a way as to be distributed, and we can extract more than one piece of this information when necessary from our memory in parallel. We are not mistaken when we say that a human brain is made up of thousands of very powerful parallel processors. It is composed of a large number of highly interconnected processing elements (neurons) working in union to solve specific problems. ANN, like people, learns by example. The chapter includes characteristics of artificial neural networks, structure of ANN, elements of artificial neural networks, pros and cons of ANN.

A review on emerging artificial intelligence (AI) techniques for air pollution forecasting: Fundamentals, application and performance