Genetic Algorithm For Parameter Optimization Research Articles

Optimization of a pyrolysis furnace using multi-jet arrays through numerical and machine learning techniques

This paper investigates the effect of multi-jet arrays on the decomposition of magnesium nitrate in a pyrolysis furnace through computational fluid dynamics (CFD) and machine learning algorithms (MLAs). The effects of temperature, arrangement, cone angle of injections and distance between pyrolysis gas inlets on the decomposition rate were studied in the CFD method. Then, regression models for decomposition rates were built using four machine learning algorithms: Neural Networks (NN), Decision Tree Regressor (DTR), Random Forest Regression (RFR), and Polynomial Regression (PR). RFR was integrated with a genetic algorithm for parameter optimization since it was judged to have the best performance after comparison. Specifically, the arrangement of gas inlets mainly affects the residence time of the particles rather than the temperature uniformity. Increasing the temperature and cone angle both favor the decomposition of magnesium nitrate in the pyrolysis furnace. The interaction of gas-solid multiple jets prolongs the residence time of particles in the pyrolysis furnace thus increasing the decomposition rate when the value of d/D is taken around 0.625. The maximum decomposition rate of 99.67% is achieved when the parameters are set to arrangement 2, temperature = 897 °C, d/D = 0.63, and cone angle = 73.8°

Joint mechanical properties estimation with a novel EMG-based knee rehabilitation robot: A machine learning approach.

Joint dynamic properties play essential roles in a wide range of biomechanical movement control. This paper develops a device with a novel mechatronic design to apply small-amplitude perturbations to the human knee. Surface Electromyography is employed to record such information; at the same time, force and position sensors collect measurements to be sent to identify human joint dynamics. For classification and estimation of force, support vector machine and support vector regression techniques are applied, respectively. We devise a genetic algorithm for parameter optimization and feature extraction within the proposed methods to improve the estimation accuracy. These are then analyzed and compared to the output of our estimation model to provide a reliable comparison. Our extensive experimental results reveal a high estimation accuracy for lower limb muscles to regulate robot impedance parameters. Although the identification method sounds similar to traditional ones, knee joint properties can be estimated by the machine learning approach from the surface Electromyography without perturbations.

Forecasting stock market price by using fuzzified Choquet integral based fuzzy measures with genetic algorithm for parameter optimization

In this paper, fuzzified Choquet integral and fuzzy-valued integrand with respect to separate measures like fuzzy measure, signed fuzzy measure and intuitionistic fuzzy measure are used to develop regression model for forecasting. Fuzzified Choquet integral is used to build a regression model for forecasting time series with multiple attributes as predictor attributes. Linear regression based forecasting models are suffering from low accuracy and unable to approximate the non-linearity in time series. Whereas Choquet integral can be used as a general non-linear regression model with respect to non classical measures. In the Choquet integral based regression model parameters are optimized by using a real coded genetic algorithm (GA). In these forecasting models, fuzzified integrands denote the participation of an individual attribute or a group of attributes to predict the current situation. Here, more generalized Choquet integral,i.e., fuzzified Choquet integral is used in case of non-linear time series forecasting models. Three different real stock exchange data are used to predict the time series forecasting model. It is observed that the accuracy of prediction models highly depends on the non-linearity of the time series.

Robust modeling for fleet assignment problem based on GASVR forecast

Fleet assignment is the essential step of overall airline flight cheduling process, and the quality of assignment strategy directly affects economy and safety of air transportation. Market demand forecast is the premise of fleet assignment, and accurate forecast is an important guarantee to reduce passenger overflow and improve aircraft utilization rate. In order to carry out scientific fleet assignment, this paper studies from two aspects: Firstly, support vector machine regression (SVR) is used to forecast flight passenger flow and solve the problem of undeterminable parameters, we present a GA-SVR model with genetic algorithm for parameter optimization. Secondly, from the perspective of flight recovery efficiency, this paper incorporates the concept of fleet robustness, and establishes the robust model for fleet assignment with dual-objective, which maximizes the flight operating profit and minimizes the number of aircraft type in busy airports. Finally, flight network of an airline is analysed to verify the validity of the model and algorithm. It shows that: the MSE mean value of GA-SVR prediction results is between 0.0103 and-0.0031, which is relatively accurate. And the fleet assignment model can significantly improve robustness (16.7%) at the expense of less profit (4.2%).

An efficient intrusion detection system based on hypergraph - Genetic algorithm for parameter optimization and feature selection in support vector machine

Realization of the importance for advanced tool and techniques to secure the network infrastructure from the security risks has led to the development of many machine learning based intrusion detection techniques. However, the benefits and limitations of these techniques make the development of an efficient Intrusion Detection System (IDS), an open challenge. This paper presents an adaptive, and a robust intrusion detection technique using Hypergraph based Genetic Algorithm (HG - GA) for parameter setting and feature selection in Support Vector Machine (SVM). Hyper – clique property of Hypergraph was exploited for the generation of initial population to fasten the search for the optimal solution and to prevent the trap at the local minima. HG-GA uses a weighted objective function to maintain the trade-off between maximizing the detection rate and minimizing the false alarm rate, along with the optimal number of features. The performance of HG-GA SVM was evaluated using NSL-KDD intrusion dataset under two scenarios (i) All features and (ii) informative features obtained from HG – GA. Experimental results show the prominence of HG-GA SVM over the existing techniques in terms of classifier accuracy, detection rate, false alarm rate, and runtime analysis.

Implementation of a Distributed Genetic Algorithm for Parameter Optimization in a Cell Nuclei Detection Project

The processing of microscopic tissue images and especially the detection of cell nuclei is nowadays done more and more using digital imagery and special immunodiagnostic software products. One of the most promising image segmentation methods is region growing, but this algorithm is very sensitive to the appropriate setting of different parameters, and the long runtime due to its high computing demand reduces its practical usability. As a result of our research, we managed to develop a data-parallel region growing algorithm that is two or three times faster than the original sequential version. The paper summarizes our results: the development of an evolution-based algorithm that was used to successfully determine a set of parameters that could be used to achieve significantly better accuracy than the already existing parameters.

Application of Heuristic Genetic Algorithm for Parameters Optimization of a Solar Cell Manufacturing Process

The more rapid development of economy, the more resources demand. The developments of solar and other alternative energy are particularly urgent and important. Diffusion manufacturing processes is core process of a solar cell manufacturing process. The physical and chemical reactions with the corresponding product characteristics are non-linear. The processes cannot be amended effectively depend on engineers experiences. In order to find the optimal process parameters for the silicon solar cell diffusion process, this research proposed an approach which combines several methods: Revised Multi-objective genetic algorithms (RMOGA) and Adaptive Multi-objective genetic algorithms (AMOGA) that integrates back-propagation neural networks (BPN), technique for order preference by similarity to ideal solution (TOPSIS), and genetic algorithms (GA) with the concept of elite sets and local search. The result of this study shows that AMOGA has the best performance to enhance the breadth and depth of the MOGA search, and also quickly convergent to the high quality and quantity optimal solutions. That provides decision-makers more diverse choices.

The Study on Robust Controller Synthesis Using Genetic Optimization Algorithm Integrating Taguchi Methods

A controller synthesis algorithm is developed in this paper. The algorithm employs the genetic algorithm for parameter optimization and Taguchi method for the planning of trails in applying the genetic algorithms. The resulting two-phase algorithm explores the orthogonal array in Taguchi method to conduct a series of experiments so that key parameters pertaining to the control factors, noise factors, and quality factors can be determined. In the first phase, a matrix-type experiment is conducted to determine the configuration for parameter optimization. The second phase then applies parameter optimization method to determine the controller parameter that leads to robust performance. The combined two-phase approach is effective and efficient in controller synthesis. The proposed algorithm is applied to a control-design benchmark problem. The resulting design is shown to have a superior performance to other existing controllers.

A new tree structure code for equivalent circuit and evolutionary estimation of parameters

To optimize the parameters of electrical elements contained in an equivalent circuit for electrochemical impedance spectroscopy, we proposed a simple, intuitive and universal tree structure code (TSC) to encode an arbitrary complex circuit, then designed a genetic algorithm for parameter optimization (GAPO) to work with the TSC and estimate the parameter values of electrical elements. The GAPO uses a novel crossover operator that performs by the non-convex linear combination of multiple parents and sets up a crossover subspace to enhance the global search. We first examined the effects of some key control parameters in the GAPO on the optimization process by selecting a relatively complex equivalent circuit to generate simulated data and comparing the parameters obtained by GAPO with the original values. Secondly, to examine the effectiveness and robustness of GAPO, we chose a set of simulated data generated by a relatively simple circuit, three sets of real impedance data on modified gold electrodes and a set of real impedance data on the anode of lithium-ion battery to run the GAPO and compared their calculated results with those obtained by complex nonlinear least square method (CNLS) supported by LEVM software. Finally, we compared the effects of five representative weighting strategies on the GAPO based on a set of simulated data generated by a relatively complicated circuit but with up to 10% Gaussian noise and the set of real impedance data on the anode of lithium-ion battery. All of these experimental results show that the GAPO works more quickly, efficiently and stably than CNLS when optimizing the element parameters. We also found that appropriate weighting strategies can help reduce the effects of experimental errors on GAPO, but the effects really depend on the nature of the specific impedance data.