Saratha Sathasivam

Design Optimization of Pin Fin Geometry Using Particle Swarm Optimization Algorithm

Particle swarm optimization (PSO) is employed to investigate the overall performance of a pin fin.The following study will examine the effect of governing parameters on overall thermal/fluid performance associated with different fin geometries, including, rectangular plate fins as well as square, circular, and elliptical pin fins. The idea of entropy generation minimization, EGM is employed to combine the effects of thermal resistance and pressure drop within the heat sink. A general dimensionless expression for the entropy generation rate is obtained by considering a control volume around the pin fin including base plate and applying the conservations equations for mass and energy with the entropy balance. Selected fin geometries are examined for the heat transfer, fluid friction, and the minimum entropy generation rate corresponding to different parameters including axis ratio, aspect ratio, and Reynolds number. The results clearly indicate that the preferred fin profile is very dependent on these parameters.

Logic mining in neural network: reverse analysis method

Neural networks are becoming very popular with data mining practitioners because they have proven through comparison their predictive power with statistical techniques using real data sets. Based on this idea, we will present a method for inducing logical rules from empirical data—Reverse Analysis. When the values of the connections of a neural network resulting from Hebbian learning for the data are given, we hope to know what logical rules are entrenched in it. This method is tested with some real life data sets. In real life data sets, logical rules are assumed to be in conjunctive normal form (CNF) since Horn clauses are inadequate.

A new hybrid quadratic regression and cascade forward backpropagation neural network

In this study, a quadratic regression model (QRM) and a cascade forward backpropagation neural network (CFBN) are jointly integrated together to form a hybrid model called the new hybrid quadratic regression method and cascade forward backpropagation neural (QRM-CFBN) network method. The new hybrid method was tested on a daily time series data obtained from the UCI repository data link and the data set was collected from a combined cycle power plant. The joint integration was made possible by the Bayesian model averaging technique, which was used to obtain a combined forecast from the two separate methods. The model resulting from the joint integration was applied on the log difference series of the original time series data. The results obtained from the new hybrid QRM-CFBN were compared with the results obtained from the hybrid ARIMA-RNN, standalone cascade forward backpropagation neural (CFBN) network and layered recurrent neural network (LRNN) after being tested on the same sample time series data respectively. The comparison indicates that the results emerging from the new hybrid QRM-CFBN method on the average, generally results in better performance when compared with the hybrid ARIMA-RNN, the standalone CFBN and the standalone LRNN for 1day, 3days as well as 5days prediction mean absolute error (MAE) and root mean square error (RMSE) for varying data samples of 50, 100, 200, 400 and 800days respectively. The RMSEs and the MAEs were applied to ascertain the assertion that the new jointly integrated forecast has better forecasting performance greater than the standalone CFBN and LRNN forecasts as well as ARIMA-RNN forecast. The analysis for this study was simulated using MATLAB software, version 8.03.

ENERGY RELAXATION FOR HOPFIELD NETWORK WITH THE NEW LEARNING RULE

In this paper, the time for energy relaxation for Little‐Hopfield neural network using the new activation rule is shown to be better than the relaxation time using Hebbian learning. However, this should be so given the characteristics of the activation function and show through computer simulations that this is indeed so. In this paper, it has been proven that the new learning rule has a higher capacity than Hebb rule by computer simulations.

Boltzmann machine and reverse analysis method

Boltzmann machine are examined for its ability to accelerate the performance of doing data mining by using technique named as Reverse Analysis method. In this paper, we describe how Hopfield network perform better with Boltzmann machine technique and able to induce logical rules from large database by using reverse analysis method: given the values of the connections of a network, we can hope to know what logical rules are entrenched in the database.

Learning in the Recurrent Hopfield Network

There are two ways to calculate synaptic weights for neurons in logic programming. There are by using Hebbian learning or by Wan Abdullahs method. Hebbian learning for governing events corresponding to some respective program clauses is equivalent with learning using Wan Abdullahs method for the same respective program clauses. We will evaluate experimentally the logical equivalent between these two types of learning (Wan Abdullahs method and Hebbian learning) for the same respective clauses (same underlying logical rules) in this paper. The computer simulation that had been carried out support this theory.

Genetic Algorithm for Restricted Maximum k-Satisfiability in the Hopfield Network

The restricted Maximum k-Satisfiability MAX- kSAT is an enhanced Boolean satisfiability counterpart that has attracted numerous amount of research. Genetic algorithm has been the prominent optimization heuristic algorithm to solve constraint optimization problem. The core motivation of this paper is to introduce Hopfield network incorporated with genetic algorithm in solving MAX-kSAT problem. Genetic algorithm will be integrated with Hopfield network as a single network. The proposed method will be compared with the conventional Hopfield network. The results demonstrate that Hopfield network with genetic algorithm outperforms conventional Hopfield networks. Furthermore, the outcome had provided a solid evidence of the robustness of our proposed algorithms to be used in other satisfiability problem.

Neuro-symbolic Performance Comparison

The convergence property for doing logic programming in Hopfield network can be accelerated by using sign constrained method and new modified learning rule. In this paper we compare the performance of doing logic programming task using the both methods. The capacity and attractor performance of these networks is tested by using computer simulations. In this paper, it has been proven by computer simulations that the sign constrained method provides better solutions.

Improving Logic Programming in Hopfield Network with Sign Constrained

The convergence property for doing logic programming in Hopfield network can be accelerated by using sign constrained method. The capacity and attractor performance of these networks is tested by using computer simulations. In this paper, it has been proven by computer simulations that the new approach provides good solutions.

On the fusion of neural network models in the case of heteroscedasticity

In this study, a neural network based clustering algorithm by robust measure was developed with the intention of filtering outliers from a given dataset. Outliers in data set usually make a model to deviate from the assumption of homoscedastic relationship which in turn leads to a heteroscedastic relationship. The developed clustering based algorithm uses two types of neural network. i.e. Cascade forward backpropagation neural network and feed forward neural network. The clustering algorithm was based on the robust estimates of location and dispersion matrix. Five (5) independent data sets obtained from the UCI machine learning repository data link were used for this study. Two techniques i.e. CFBNFDCARM and FFNNFDCARM were employed with the aim of comparing the performance of each of the technique. The evaluating metrics of mean squared error (MSE), root mean square error (RMSE), mean absolute error (MAE) and the mean absolute percentage error (MAPE) were used as the measure of the performance function in this study. From the obtained results, CFBNFDCARM technique came out to be better than the FFNNFDCARM technique. The analyses of this study were simulated using both MATLAB R2014 software and R i386 version 3.3.0 software. The performance of each technique were further plotted in order to illustrate the findings of this study.In this study, a neural network based clustering algorithm by robust measure was developed with the intention of filtering outliers from a given dataset. Outliers in data set usually make a model to deviate from the assumption of homoscedastic relationship which in turn leads to a heteroscedastic relationship. The developed clustering based algorithm uses two types of neural network. i.e. Cascade forward backpropagation neural network and feed forward neural network. The clustering algorithm was based on the robust estimates of location and dispersion matrix. Five (5) independent data sets obtained from the UCI machine learning repository data link were used for this study. Two techniques i.e. CFBNFDCARM and FFNNFDCARM were employed with the aim of comparing the performance of each of the technique. The evaluating metrics of mean squared error (MSE), root mean square error (RMSE), mean absolute error (MAE) and the mean absolute percentage error (MAPE) were used as the measure of the performance function i...