Trapping In Local Minima Research Articles

Genetic optimization of neural network and fuzzy logic for oil bubble point pressure modeling

Bubble point pressure is a critical pressure-volume-temperature (PVT) property of reservoir fluid, which plays an important role in almost all tasks involved in reservoir and production engineering. We developed two sophisticated models to estimate bubble point pressure from gas specific gravity, oil gravity, solution gas oil ratio, and reservoir temperature. Neural network and adaptive neuro-fuzzy inference system are powerful tools for extracting the underlying dependency of a set of input/output data. However, the mentioned tools are in danger of sticking in local minima. The present study went further by optimizing fuzzy logic and neural network models using the genetic algorithm in charge of eliminating the risk of being exposed to local minima. This strategy is capable of significantly improving the accuracy of both neural network and fuzzy logic models. The proposed methodology was successfully applied to a dataset of 153 PVT data points. Results showed that the genetic algorithm can serve the neural network and neurofuzzy models from local minima trapping, which might occur through back-propagation algorithm.

A novel hybrid genetic differential evolution algorithm for constrained optimization problems

Most of the real-life applications have many constraints and they are considered as constrained optimization problems (COPs). In this paper, we present a new hybrid genetic differential evolution algorithm to solve constrained optimization problems. The proposed algorithm is called hybrid genetic differential evolution algorithm for solving constrained optimization problems (HGDESCOP). The main purpose of the proposed algorithm is to improve the global search ability of the DE algorithm by combining the genetic linear crossover with a DE algorithm to explore more solutions in the search space and to avoid trapping in local minima. In order to verify the general performance of the HGDESCOP algorithm, it has been compared with 4 evolutionary based algorithms on 13 benchmark functions. The experimental results show that the HGDESCOP algorithm is a promising algorithm and it outperforms other algorithms.

Improved Artificial Fish Algorithm for Parameters Optimization of PID Neural Network

In order to solve problems such as initial weights are difficult to be determined, training results are easy to trap in local minima in optimization process of PID neural network parameters by traditional BP algorithm, this paper proposed a new method based on improved artificial fish algorithm for parameters optimization of PID neural network. This improved artificial fish algorithm uses a composite adaptive artificial fish algorithm based on optimal artificial fish and nearest artificial fish to train network weights parameters of PID neural network. By comparing food consistence in preying behavior to adaptively select vision and step of artificial fish, this method overcomes shortcomings such as slow convergence speed, low optimization accuracy of basic artificial fish algorithm. Simulations of PID neural network system whose parameters are trained respectively through BP algorithm and improved artificial fish algorithm are conducted respectively in the MATLAB environment. The simulation result shows that the PID neural network control system whose parameters are trained by the improved artificial fish algorithm has a better control effect, especially for nonlinear systems

Combined ANN Prediction Model for Rock Fragmentation Distribution due to Blasting

It is feasible using Artificial Neural Network (ANN) to predict the distribution of particle size in bench blasting. It explored the minimum error based on the steepest descent method and plunged into local minima easily. A Back-propagation Neural Network (BPNN) improved by genetic algorithm (GA-BP) to predict mean particle size (K50) was proposed. The input variables used for the proposed method were rock joints (J), Protogyakonov’s coefficient (f ), burden (W ), time interval of millisecond (T ), detonation velocity (V ), explosive specific charge (G), stemming length (L) and hole-space (S). After using the practical data for training and testing, the forecasted mean Relative Errors (RE) of Multivariate Regression Analysis (MVRA), BPNN were 11:15% and 7:38% respectively while the RE achieved by GA-BP was 3:09%. Results indicate that GA-BP model can efficiently reach the expected target, and the problem of incidental trap in local minimum is solved. The improved model is more suitable for prediction of blasting fragmentation.

An intelligent approach for optimal prediction of gas deviation factor using particle swarm optimization and genetic algorithm

The measurement of PVT properties of natural gas in gas pipelines, gas storage systems, and gas reservoirs require accurate values of compressibility factor. Although equation of state and empirical correlations were utilized to estimate compressibility factor, but the demands for novel, more reliable, and easy-to-use models encouraged the researchers to introduce modern tools such as artificial intelligent systems.This paper introduces Particle swarm optimization (PSO) and Genetic algorithm (GA) as population-based stochastic search algorithms to optimize the weights and biases of networks, and to prevent trapping in local minima. Hence, in this paper, GA and PSO were used to minimize the neural network error function.A database containing 6378 data was employed to develop the models. The proposed models were compared to conventional correlations so that the model predictions indicated a good accuracy for the results in training and testing stages. The results showed that artificial neural networks (ANNs) remarkably overcame the inadequacies of the empirical models where PSO–ANN improved the performance significantly. Additionally, the regression analysis released the efficiency coefficient (R2) of 0.999 which can be considered very promising.

Simultaneous Inversion for Velocity and Reflector Geometry Using Multi-phase Fresnel Volume Rays

Traditional ray tomography methods based on the high frequency assumption are sometimes unable to obtain a high resolution tomographic picture due to a deficient coverage of ray paths in real applications, especially for low velocity anomalous regions. In contrast, finite-frequency ray theory is more suitable for handling real seismic propagation problems because the travel time depends not only on the velocity distribution along a central ray (or traditional geometric ray), but also on the velocity values within a region (referred to as the first Fresnel Volume) which incorporates the central ray. In this study, we develop an algorithm to calculate multi-phase Fresnel Volume finite-frequency rays, and then present an inversion method to simultaneous invert for both velocity and reflector geometry by using these multi-phase Fresnel Volume finite-frequency rays. Using synthetic data examples, we compare the reconstructions of the velocity field and the reflector orientation using the Fresnel Volume ray tomographic methods and the traditional ray tomography approach. Results show that the former is advantageous over the latter, especially when the ray density is relatively low. An additional benefit of the Fresnel Volume finite-frequency ray tomographic method is that it can start with a low frequency to capture the coarse velocity structure, thereby mitigating the local minimum trapping problem, and then be tuned to a high frequency for delineating the fine velocity structure.

Segmentation and Edge Detection Based on Modified ant Colony Optimization for Iris Image Processing

Abstract Ant colony optimization (stocktickerACO) is a meta-heuristic algorithm inspired by food searching behavior of real ants. Recently stocktickerACO has been widely used in digital image processing. When artificial ants move in a discrete habitat like an image, they deposit pheromone in their prior position. Simultaneously, vaporizing of pheromone in each iteration step avoids from falling in the local minima trap. Iris recognition because of its great dependability and non-invasion has various applications. simulation results demonstrate stocktickerACO algorithm can effectively extract the iris texture. Also it is not sensitive to nuisance factors. Moreover, stocktickerACO in this research preserves details of the various synthetic and real images. Performance of ACO in iris segmentation is compared with operation of traditional approaches such as canny, robert, and sobel edge detections. Experimental results reveal high quality and quite promising of stocktickerACO to segment images with irregular and complex structures.

有考虑负载不确定性的多目标最佳虚功率调度

由于传统的最佳虚功率调度都是将负载需求假设为已知的固定值，以求得负载在确定状态下的最佳解，但实际的负载需求含有不确定性。为了使调度结果更适用于实际情形，本文探讨考虑负载不确定性的多目标最佳虚功率调度问题。此外，本文提出以增强型萤火虫演算法(Enhanced Firefly Algorithm, EFA)应用于此问题，增强型萤火虫演算法是将原萤火虫演算法的更新式作改良且修改参数并加入突变机制，以增强演算法的开采与搜索能力，并加快收敛速度且不易陷入局部解。另外，本文使用模糊理论建立模糊归属函数，以解决多目标不同性质及目标函数“越小越好”这种不明确语意的问题。为了验证本文所使用的方法对于考虑负载不确定性的多目标最佳虚功率调度问题的有效性，本文使用IEEE57-Bus系统作为测试系统，并与其它演算法作比较。实验结果证实本文所提出的方法确实可以获得较好的结果。 In traditional optimal reactive power dispatch problem, the optimal solution is found under the condition that the load demands are assumed to be known and fixed, but the practical load demands have uncertainty. This paper investigates the multi-objective optimal reactive power dispatch with considering load uncertainty to make the dispatch results more suitable for real situation. In this paper, an enhanced firefly algorithm is presented to solve the problem. Enhanced firefly algorithm is based on firefly algorithm that the update formula and parameters are modified and the mutation strategy is utilized to enhance the capabilities of exploring and searching. So the proposed algorithm can converge fast and the solution can avoid trapping in local minimum. Furthermore, in order to deal with the multi-objective problem and the ambiguous linguistic expression such as “as little as possible”, the fuzzy theory is employed to establish the fuzzy membership functions. To demonstrate the effectiveness of the proposed method for solving multi-objective optimal reactive power dispatch with considering load uncertainty problem, the IEEE 57-Bus system has been applied to the reactive power dispatching and the results of the proposed method are compared with those of other algorithms. The results show that the proposed method can get better solution.

Towards Behavior Control for Evolutionary Robot Based on RL with ENN

This paper proposes a behavior-switching control strategy of an evolutionary robot based on Artificial Neural Network (ANN) and genetic algorithm (GA). This method is able not only to construct the reinforcement learning models for autonomous robots and evolutionary robot modules that control behaviors and reinforcement learning environments, and but also to perform the behavior-switching control and obstacle avoidance of an evolutionary robot in the unpredictable environments with the static and moving obstacles by combining ANN and GA. The experimental results have demonstrated that our method can perform the decision-making strategy and parameter setup optimization of ANN and GA by learning and can effectively escape from a trap of local minima, avoid motion deadlock status of humanoid soccer robotic agents, and reduce the oscillation of the planned trajectory among the multiple obstacles by crossover and mutation. We have successfully applied some results of the proposed algorithm to our simulation humanoid robotic soccer team CIT3D which won the 1st prize of RoboCup Championship and ChinaOpen2010 and the 2nd place of the official RoboCup World Championship on 5-11, July 2011 in Istanbul, Turkey. In comparison with the conventional behavior network and the adaptive behavior method, our algorithm simplified the genetic encoding complexity, improved the convergence rate ρ and the network performance.

Using Artificial Bee Colony to Improve Functional Link Neural Network Training

Artificial Neural Networks have emerged as an important tool for classification and have been widely used to classify non-linearly separable pattern. The most popular artificial neural networks model is a Multilayer Perceptron (MLP) that is able to perform classification task with significant success. However due to the complexity of MLP structure and also problems such as local minima trapping, over fitting and weight interference have made neural network training difficult. Thus, the easy way to avoid these problems is by removing the hidden layers. This paper presents the ability of Functional Link Neural Network (FLNN) in overcoming the complexity structure of MLP, using it single layer architecture and proposes an Artificial Bee Colony (ABC) optimization for training the FLNN. The proposed technique is expected to provide better learning scheme for a classifier in order to get more accurate classification result.

Integrating piecewise linear representation and weighted support vector machine for stock trading signal prediction

Piecewise linear representation (PLR) and back-propagation artificial neural network (BPN) have been integrated for the stock trading signal prediction recently (PLR–BPN). However, there are some disadvantages in avoiding over-fitting, trapping in local minimum and choosing the threshold of the trading decision. Since support vector machine (SVM) has a good way to avoid over-fitting and trapping in local minimum, we integrate PLR and weighted SVM (WSVM) to forecast the stock trading signals (PLR–WSVM). The new characteristics of PLR–WSVM are as follows: (1) the turning points obtained from PLR are set by different weights according to the change rate of the closing price between the current turning point and the next one, in which the weight reflects the relative importance of each turning point; (2) the prediction of stock trading signal is formulated as a weighted four-class classification problem, in which it does not need to determine the threshold of trading decision; (3) WSVM is used to model the relationship between the trading signal and the input variables, which improves the generalization performance of prediction model; (4) the history dataset is divided into some overlapping training–testing sets rather than training–validation–testing, which not only makes use of data fully but also reduces the time variability of data; and (5) some new technical indicators representing investors’ sentiment are added to the input variables, which improves the prediction performance. The comparative experiments among PLR–WSVM, PLR–BPN and buy-and-hold strategy (BHS) on 20 shares from Shanghai Stock Exchange in China show that the prediction accuracy and profitability of PLR–WSVM are all the best, which indicates PLR–WSVM is effective and can be used in the stock trading signal prediction.

Single Neuron PID Adaptive Control for BLDCM Based on Algebraic Algorithm

The brushless DC motor control system often adopts the classic PID control, the advantages of which are as follows: simple to control, easy to adjust the parameter and a certain degree of control precision. But it relies on accurate mathematical model. The permanent magnet brushless DC motor control system is a multi-variable and nonlinear. As to the deficiencies of the classic PID control method, this thesis proposes a method called artificial neural network PID adaptive control method, which is based on algebraic algorithm and overcomes the shortcomings such as the slow convergence of BP algorithm, easy to trap in local minimum, and etc.

G-HABC Algorithm for Training Artificial Neural Networks

Learning problems for Neural Network (NN) has widely been explored in the past two decades. Researchers have focused more on population-based algorithms because of its natural behavior processing. The population-based algorithms are Ant Colony Optimization (ACO), Artificial Bee Colony (ABC), and recently Hybrid Ant Bee Colony (HABC) algorithm produced an easy way for NN training. These social based techniques are mostly used for finding best weight values and over trapping local minima in NN learning. Typically, NN trained by traditional approach, namely the Backpropagation (BP) algorithm, has difficulties such as trapping in local minima and slow convergence. The new method named Global Hybrid Ant Bee Colony (G-HABC) algorithm which can overcome the gaps in BP is used to train the NN for Boolean Function classification task. The simulation results of the NN when trained with the proposed hybrid method were compared with that of Levenberg-Marquardt (LM) and ordinary ABC. From the results, the proposed G-HABC algorithm has shown to provide a better learning performance for NNs with reduced CPU time and higher success rates.

Training feedforward neural networks using hybrid particle swarm optimization and gravitational search algorithm

The Gravitational Search Algorithm (GSA) is a novel heuristic optimization method based on the law of gravity and mass interactions. It has been proven that this algorithm has good ability to search for the global optimum, but it suffers from slow searching speed in the last iterations. This work proposes a hybrid of Particle Swarm Optimization (PSO) and GSA to resolve the aforementioned problem. In this paper, GSA and PSOGSA are employed as new training methods for Feedforward Neural Networks (FNNs) in order to investigate the efficiencies of these algorithms in reducing the problems of trapping in local minima and the slow convergence rate of current evolutionary learning algorithms. The results are compared with a standard PSO-based learning algorithm for FNNs. The resulting accuracy of FNNs trained with PSO, GSA, and PSOGSA is also investigated. The experimental results show that PSOGSA outperforms both PSO and GSA for training FNNs in terms of converging speed and avoiding local minima. It is also proven that an FNN trained with PSOGSA has better accuracy than one trained with GSA.

Optimal Allocation and Sizing of Active Power Line Conditioners Using a New Particle Swarm Optimization-based Approach

Optimal allocation and sizing of active power line conditioners in a distribution network is a discrete and non-linear problem. Heuristic-based optimization methods are reliable approaches for solving these types of problems, particularly for a large-scale distribution network. However, the optimization parameters in heuristic methods can influence the final solution significantly so that if these parameters are selected improperly, the risk of trapping in local minima may be remarkable. In this article, a new method—modified particle swarm optimization—is proposed for optimal planning of active power line conditioners in a distribution network that is polluted by non-linear loads. In this method, external particle swarm optimization is used to determine the main particle swarm optimization parameters accurately. To validate the performance of the proposed method, a 6-bus and the IEEE 18-bus test systems are employed. Different scenarios are studied, and the results are compared with those obtained by other optimization methods, such as conventional particle swarm optimization, genetic algorithm, and discrete non-linear programming. The results illustrate higher accuracy of the proposed modified particle swarm optimization and its applicability for solving discrete and non-linear problems, such as active power line conditioner planning.

Of possible cheminformatics futures

For over a decade, cheminformatics has contributed to a wide array of scientific tasks from analytical chemistry and biochemistry to pharmacology and drug discovery; and although its contributions to decision making are recognized, the challenge is how it would contribute to faster development of novel, better products. Here we address the future of cheminformatics with primary focus on innovation. Cheminformatics developers often need to choose between "mainstream" (i.e., accepted, expected) and novel, leading-edge tools, with an increasing trend for open science. Possible futures for cheminformatics include the worst case scenario (lack of funding, no creative usage), as well as the best case scenario (complete integration, from systems biology to virtual physiology). As "-omics" technologies advance, and computer hardware improves, compounds will no longer be profiled at the molecular level, but also in terms of genetic and clinical effects. Among potentially novel tools, we anticipate machine learning models based on free text processing, an increased performance in environmental cheminformatics, significant decision-making support, as well as the emergence of robot scientists conducting automated drug discovery research. Furthermore, cheminformatics is anticipated to expand the frontiers of knowledge and evolve in an open-ended, extensible manner, allowing us to explore multiple research scenarios in order to avoid epistemological "local information minimum trap".

New Potential Functions with Random Force Algorithms Using Potential Field Method

Autonomous mobile robot path planning is a common topic for robotics and computational geometry. Many important results have been found, but a lot of issues are still veiled. This paper first describes new problem of symmetrically aligned robot-obstacle-goal (SAROG) when using potential field methods for mobile robot path planning. In addition, we consider constant robot speed for practical use. The SAROG and the constant speed involve two potential risks: robot-obstacle collision and local minima trap. For dealing with the two potential risks, we analyze the conditions of the collision and the local minima trap, and propose new potential functions and random force based algorithms. For the algorithm verification, we use WiRobot X80 with three ultrasonic range sensor modules.

Artificial Bee Colony Optimization of Capping Potentials for Hybrid Quantum Mechanical/Molecular Mechanical Calculations.

We present an algorithmic extension of a numerical optimization scheme for analytic capping potentials for use in mixed quantum-classical (quantum mechanical/molecular mechanical, QM/MM) ab initio calculations. Our goal is to minimize bond-cleavage-induced perturbations in the electronic structure, measured by means of a suitable penalty functional. The optimization algorithm-a variant of the artificial bee colony (ABC) algorithm, which relies on swarm intelligence-couples deterministic (downhill gradient) and stochastic elements to avoid local minimum trapping. The ABC algorithm outperforms the conventional downhill gradient approach, if the penalty hypersurface exhibits wiggles that prevent a straight minimization pathway. We characterize the optimized capping potentials by computing NMR chemical shifts. This approach will increase the accuracy of QM/MM calculations of complex biomolecules.

A New Neural Network Approach to Short Term Load Forecasting of Electrical Power Systems

Short-term load forecast (STLF) is an important operational function in both regulated power systems and deregulated open electricity markets. However, STLF is not easy to handle due to the nonlinear and random-like behaviors of system loads, weather conditions, and social and economic environment variations. Despite the research work performed in the area, more accurate and robust STLF methods are still needed due to the importance and complexity of STLF. In this paper, a new neural network approach for STLF is proposed. The proposed neural network has a novel learning algorithm based on a new modified harmony search technique. This learning algorithm can widely search the solution space in various directions, and it can also avoid the overfitting problem, trapping in local minima and dead bands. Based on this learning algorithm, the suggested neural network can efficiently extract the input/output mapping function of the forecast process leading to high STLF accuracy. The proposed approach is tested on two practical power systems and the results obtained are compared with the results of several other recently published STLF methods. These comparisons confirm the validity of the developed approach.

The Risk Element Transmission Model of Power Restoration Based on Dynamic Improved Ant Colony Algorithm

In order to restore power supply of a complex distribution network in the most economical way, proposed a dynamic improved ant colony algorithm. It’s used to select the path of supply restoration. Improved the original pheromone update mechanism of ant colony algorithm and update the increment of pheromone dynamically. It not only takes into account the inherent loss in the chosen route for power restoration, but also considers the economic losses in every region caused by time delay. This algorithm can find approximate extreme solutions quickly and avoid trapping in local minima. It can minimum the total economic loss of the electric interruption, and provide a reliable basis for the economic decision-making of the power recovery.