Improved monarch butterfly optimization for unconstrained global search and neural network training

This work is a seminal attempt to address the drawbacks of the recently proposed monarch butterfly optimization (MBO) algorithm. This algorithm suffers from premature convergence, which makes it less suitable for solving real-world problems. The position updating of MBO is modified to involve previous solutions in addition to the best solution obtained thus far. To prove the efficiency of the Improved MBO (IMBO), a set of 23 well-known test functions is employed. The statistical results show that IMBO benefits from high local optima avoidance and fast convergence speed which helps this algorithm to outperform basic MBO and another recent variant of this algorithm called greedy strategy and self-adaptive crossover operator MBO (GCMBO). The results of the proposed algorithm are compared with nine other approaches in the literature for verification. The comparative analysis shows that IMBO provides very competitive results and tends to outperform current algorithms. To demonstrate the applicability of IMBO at solving challenging practical problems, it is also employed to train neural networks as well. The IMBO-based trainer is tested on 15 popular classification datasets obtained from the University of California at Irvine (UCI) Machine Learning Repository. The results are compared to a variety of techniques in the literature including the original MBO and GCMBO. It is observed that IMBO improves the learning of neural networks significantly, proving the merits of this algorithm for solving challenging problems.     

A multi-verse optimizer approach for feature selection and optimizing SVM parameters based on a robust system architecture

Neural Computing and Applications - Support vector machine (SVM) is a well-regarded machine learning algorithm widely applied to classification tasks and regression problems. SVM was founded based...     

Increased In Vitro and In Vivo Digestibility of Soy Proteins by Chemical Modification of Disulfide Bonds

To improve protein digestibility of aqueously extracted soy proteins, an effective chemical treatment under mild conditions is needed. Soy proteins, including storage protein glycinin and antinutritional factors such as trypsin inhibitors, are rich in disulfide bonds. Reduction of these disulfide bonds by incubating soy proteins with sodium sulfite and sodium metabisulfite at 55 °C showed no net increase of free sulfhydryl groups after dialysis to remove the residual reducing agent. However, the in vitro digestibility measured by trypsin hydrolysis using the pH-Stat method was significantly increased. Sodium metabisulfite (SMBS) was more effective in increasing in vitro digestibility than sodium sulfite at the same molar concentration. The digestibility of soy protein treated by 0.5 mmol SMBS/g soy flour at 55 °C was more than doubled compared to that of the control without reduction treatment. Large-scale testing of soy proteins treated with SMBS for an in vivo animal feeding study showed similar in vitro digestibility by trypsin, e.g., the degree of hydrolysis of the treated sample was 8.5% compared to 1.6% of the control. These soy proteins were further evaluated using a chick growth model. The protein efficiency ratio (PER) increased by 57% when the chicks fed SMBS-treated soy were compared to the chicks fed raw soy flour. SMBS-fed chicks did not display any pancreatic hypertrophy compared to those fed with raw soy control. These results indicate that there is great potential to use safe chemicals and mild temperature to inactivate the antinutritional factors in soybeans and thus improve digestibility of soy proteins that are extracted with low-temperature aqueous process.     

Compatibilization of low‐density polyethylene/plasticized starch blends by reactive compounds and electron beam irradiation

Blends based on various compositions of low‐density polyethylene (LDPE) and plasticized starch (PLST) were prepared by melt extrusion and molding in the form of sheets under hot press. The rheology properties during mixing were studied in terms of torque and temperature against mixing time. The structural properties of LDPE/PLST blends before and after electron beam irradiation was characterized by IR spectroscopy, tensile mechanical testing, and scanning electron microscopy (SEM). The torque‐time curves during the mixing process showed that the values of torque in the first region of mixing for pure LDPE or LDPE/PLST blends are higher in the presence of the compatibilizer PEMA than that in the presence of EVA. In addition, the stability of mixing was attained after a short time in the presence of PEMA. The IR spectroscopy suggests that the compatibilization by EVA and PEMA compounds proceeds through the formation of hydrogen bonding during mixing and this compatibility was improved after electron beam irradiation. The stress–strain curves of pure LDPE and its blends with PLST showed the behavior of tough polymers with yielding properties. The SEM micrographs of the fracture surfaces give supports to the effect of EVA and PEMA as compatibilizers and the effect of electron beam irradiation. POLYM. COMPOS., 2008. © 2008 Society of Plastics Engineers     

Microstructure Characterization and Corrosion Resistance Behavior of New Cobalt-Free Maraging Steel Produced Through ESR Techniques

In this study, two different grades (M23 and M29) of cobalt-free low nickel maraging steel have been produced through electroslag remelting (ESR) process. The corrosion resistance of these ESR steels was investigated in 1 M H₂SO₄ solution using linear potentiodynamic polarization (LPP) and electrochemical impedance spectroscopy (EIS) techniques. The experiments were performed for different immersion time and solution temperature. To evaluate the corrosion resistance of the ESR steels, some significant characterization parameters from LPP and EIS curves were analyzed and compared with that of conventional C250 maraging steel. Irrespective of measurement techniques used, the results show that the corrosion resistance of the ESR steels was higher than the C250 steel. The microstructure of ESR steels was composed of uniform and well-distributed martensite accompanied with little amount of retained austenite in comparison with C250 steel.     

Experimentally validated analytical modeling of diesel engine power and in-cylinder gas speed dynamics

Supercharged diesel engines are a key element in diesel powertrains that have been extensively modelled yet often without explainable mathematical trends. The present paper demonstrates the analytical modelling of in-cylinder gas speed dynamics and engine brake power. These analytical models provide explainable mathematical trends. In addition, they provide gear-shifting-based modeling because the model parameters can be adjusted to reflect different driving conditions without the need for gathering field data. An unprecedented sensitivity analysis was conducted on these developed models for simplifying them. They were validated using experimental data and the relative error of the developed model of the in-cylinder gas speed dynamics was 9.8%. The study demonstrates with 73% coefficient of determination that the average percentage of deviation of the simulated results from the corresponding field data on the engine brake power is 6.9%. The relative error of the developed model of the engine brake power is 7%. These values of relative error are an order of magnitude of deviation that is less than that of widely recognized models in the field of vehicle powertrain modeling such as the CMEM and GT-Power. These analytically developed models serve as widely valid models. Having addressed and corrected flaws in the corresponding models, such as the model of the in-cylinder gas speed dynamics presented in a key reference in this research area, these developed models can help in better analyzing and assessing the performance of diesel engines.

     

Recognizing arabic handwritten characters using deep learning and genetic algorithms

Automated techniques for Arabic content recognition are at a beginning period contrasted with their partners for the Latin and Chinese contents recognition. There is a bulk of handwritten Arabic archives available in libraries, data centers, historical centers, and workplaces. Digitization of these documents facilitates (1) to preserve and transfer the country’s history electronically, (2) to save the physical storage space, (3) to proper handling of the documents, and (4) to enhance the retrieval of information through the Internet and other mediums. Arabic handwritten character recognition (AHCR) systems face several challenges including the unlimited variations in human handwriting and the leakage of large and public databases. In the current study, the segmentation and recognition phases are addressed. The text segmentation challenges and a set of solutions for each challenge are presented. The convolutional neural network (CNN), deep learning approach, is used in the recognition phase. The usage of CNN leads to significant improvements across different machine learning classification algorithms. It facilitates the automatic feature extraction of images. 14 different native CNN architectures are proposed after a set of try-and-error trials. They are trained and tested on the HMBD database that contains 54,115 of the handwritten Arabic characters. Experiments are performed on the native CNN architectures and the best-reported testing accuracy is 91.96%. A transfer learning (TF) and genetic algorithm (GA) approach named “HMB-AHCR-DLGA” is suggested to optimize the training parameters and hyperparameters in the recognition phase. The pre-trained CNN models (VGG16, VGG19, and MobileNetV2) are used in the later approach. Five optimization experiments are performed and the best combinations are reported. The highest reported testing accuracy is 92.88%.

     

Integration of PV system with SMES based on model predictive control for utility grid reliability improvement.

This paper describes the integration of a photovoltaic (PV) renewable energy source with a superconducting magnetic energy storage (SMES) system. The integrated system can improve the voltage stability of the utility grid and achieve power leveling. The control schemes employ model predictive control (MPC), which has gained significant attention in recent years because of its advantages such as fast response and simple implementation. The PV system provides maximum power at various irradiation levels using the incremental conductance technique (INC). The interfaced grid side converter of the SMES can control the grid voltage by regulating its injected reactive power to the grid, while the charge and discharge operation of the SMES coil can be managed by the system operator to inject/absorb active power to/from the grid to achieve the power leveling strategy. Simulation results based on MATLAB/Simulink® software prove the fast response of the system control objectives in tracking the setpoints at different loading scenarios and PV irradiance levels, while the SMES injects/absorbs active and reactive power to/from the grid during various events to improve the voltage response and achieve power leveling strategy.