Parameters selection in gene selection using Gaussian kernel support vector machines by genetic algorithm

In microarray-based cancer classification, gene selection is an important issue owing to the large number of variables and small number of samples as well as its non-linearity. It is difficult to get satisfying results by using conventional linear sta- tistical methods. Recursive feature elimination based on support vector machine (SVM RFE) is an effective algorithm for gene selection and cancer classification, which are integrated into a consistent framework. In this paper, we propose a new method to select parameters of the aforementioned algorithm implemented with Gaussian kernel SVMs as better alternatives to the common practice of selecting the apparently best parameters by using a genetic algorithm to search for a couple of optimal parameter. Fast implementation issues for this method are also discussed for pragmatic reasons. The proposed method was tested on two repre- sentative hereditary breast cancer and acute leukaemia datasets. The experimental results indicate that the proposed method per- forms well in selecting genes and achieves high classification accuracies with these genes.

FPGA PLACEMENT OPTIMIZATION BY TWO-STEP UNIFIED GENETIC ALGORITHM AND SIMULATED ANNEALING ALGORITHM

Genetic Algorithm (GA) is a biologically inspired technique and widely used to solve numerous combinational optimization problems. It works on a population of individuals, not just one single solution. As a result, it avoids converging to the local optimum. However, it takes too much CPU time in the late process of GA. On the other hand, in the late process Simulated Annealing (SA) converges faster than GA but it is easily trapped to local optimum. In this letter, a useful method that unifies GA and SA is introduced, which utilizes the advantage of the global search ability of GA and fast convergence of SA. The experimental results show that the proposed algorithm outperforms GA in terms of CPU time without degradation of performance. It also achieves highly comparable placement cost compared to the state-of-the-art results obtained by Versatile Place and Route (VPR) Tool.

New Antenna Array Beamforming Techniques Based on Hybrid Convolution/Genetic Algorithm for 5G and Beyond Communications

Side lobe level reduction(SLL)of antenna arrays significantly enhances the signal-to-interference ratio and improves the quality of service(QOS)in recent and future wireless communication systems starting from 5G up to 7G.Furthermore,it improves the array gain and directivity,increasing the detection range and angular resolution of radar systems.This study proposes two highly efficient SLL reduction techniques.These techniques are based on the hybridization between either the single convolution or the double convolution algorithms and the genetic algorithm(GA)to develop the Conv/GA andDConv/GA,respectively.The convolution process determines the element’s excitations while the GA optimizes the element spacing.For M elements linear antenna array(LAA),the convolution of the excitation coefficients vector by itself provides a new vector of excitations of length N=(2M−1).This new vector is divided into three different sets of excitations including the odd excitations,even excitations,and middle excitations of lengths M,M−1,andM,respectively.When the same element spacing as the original LAA is used,it is noticed that the odd and even excitations provide a much lower SLL than that of the LAA but with amuch wider half-power beamwidth(HPBW).While the middle excitations give the same HPBWas the original LAA with a relatively higher SLL.Tomitigate the increased HPBWof the odd and even excitations,the element spacing is optimized using the GA.Thereby,the synthesized arrays have the same HPBW as the original LAA with a two-fold reduction in the SLL.Furthermore,for extreme SLL reduction,the DConv/GA is introduced.In this technique,the same procedure of the aforementioned Conv/GA technique is performed on the resultant even and odd excitation vectors.It provides a relatively wider HPBWthan the original LAA with about quad-fold reduction in the SLL.

Geochemical Characteristics and Genetic Types of Natural Gas in the Xinchang Gas Field, Sichuan Basin, SW China

The molecular compositions and stable carbon and hydrogen isotopic compositions of natural gas from the Xinchang gas field in the Sichuan Basin were investigated to determine the genetic types. The natural gas is mainly composed of methane （88.99%-98.01%）, and the dryness coefficient varies between 0.908 and 0.997. The gas generally displays positive alkane carbon and hydrogen isotopic series. The geochemical characteristics and gas-source correlation indicate that the gases stored in the 5th member of the Upper Triassic Xujiahe Formation are coal-type gases which are derived from source rocks in the stratum itself. The gases reservoired in the 4th member of the Xujiahe Formation and Jurassic strata in the Xinchang gas field are also coal-type gases that are derived from source rocks in the 3rd and 4th members of the Xujiahe Formation. The gases reservoired in the 2nd member of the Upper Triassic Xujiahe Formation are mainly coal-type gases with small amounts of oil-type gas that is derived from source rocks in the stratum itself. This is accompanied by a small amount of contribution brought by source rocks in the Upper Triassic Ma＇antang and Xiaotangzi formations. The gases reservoired in the 4th member of the Middle Triassic Leikoupo Formation are oil-type gases and are believed to be derived from the secondary cracking of oil which is most likely to be generated from the Upper Permian source rocks.

Design of artificial neural networks using a genetic algorithm to predict saturates of vacuum gas oil

Accurate prediction of chemical composition of vacuum gas oil （VGO） is essential for the routine operation of refineries. In this work, a new approach for auto-design of artificial neural networks （ANN） based on a genetic algorithm （GA） is developed for predicting VGO saturates. The number of neurons in the hidden layer, the momentum and the learning rates are determined by using the genetic algorithm. The inputs for the artificial neural networks model are five physical properties, namely, average boiling point, density, molecular weight, viscosity and refractive index. It is verified that the genetic algorithm could find the optimal structural parameters and training parameters of ANN. In addition, an artificial neural networks model based on a genetic algorithm was tested and the results indicated that the VGO saturates can be efficiently predicted. Compared with conventional artificial neural networks models, this approach can improve the prediction accuracy.

Genetic types and distribution of shallow-buried natural gases

Great volumes of shallow-buried （〈2,000 m） natural gases which are mainly composed of biogases and low-mature gases have been found in the Mesozoic-Cenozoic sedimentary basins in China. Many shallow gas reservoirs in China are characterized by coexistence of biogas and low-mature gas, so identifying the genetic types of shallow gases is important for exploration and development in sedimentary basins. In this paper, we study the gas geochemistry characteristics and distribution in different basins, and classify the shallow gas into two genetic types, biogas and low-mature gas. The biogases are subdivided further into two subtypes by their sources, the source rock-derived biogas and hydrocarbon-derived biogas. Based on the burial history of the source rocks, the source rock-derived biogases are divided into primary and secondary biogas. The former is generated from the source rocks in the primary burial stage, and the latter is from uplifted source rocks or those in a secondary burial stage. In addition, the identifying parameters of each type of shallow gas are given. Based on the analysis above, the distributions of each type of shallow gas are studied. The primary biogases generated from source rocks are mostly distributed in Quaternary basins or modem deltas. Most of them migrate in watersoluble or diffused mode, and their migration distance is short. Reservoir and caprock assemblages play an important role in primary biogas accumulation. The secondary biogases are distributed in a basin with secondary burial history. The oil-degraded biogases are distributed near heavy oil pools. The low-mature gases are widely distributed in shallow-buried reservoirs in the Meso-Cenozoic basins.

The Genetic Mechanism and Model of Deep-Basin Gas Accumulation and Methods for Predicting the Favorable Areas

As a kind of abnormal natural gas formed with special mechanism, the deep-basin gas, accumulated in the lower parts of a basin or syncline and trapped by a tight reservoir, has such characteristics as gas-water inversion, abnormal pressure, continuous distribution and tremendous reserves. Being a geological product of the evolution of petroliferous basins by the end of the middle-late stages, the formation of a deep-basin gas accumulation must meet four conditions, i.e., continuous and sufficient gas supply, tight reservoirs in continuous distribution, good sealing caps and stable structures. The areas, where the expansion force of natural gas is smaller than the sum of the capillary force and the hydrostatic pressure within tight reservoirs, are favorable for forming deep-basin gas pools. The range delineated by the above two forces corresponds to that of the deep-basin gas trap. Within the scope of the deep-basin gas trap, the balance relationship between the amounts of ingoing and overflowing gases determines the gas-bearing area of the deep-basin gas pool. The gas volume in regions with high porosity and high permeability is worth exploring under current technical conditions and it is equivalent to the practical resources (about 10%-20% of the deep-basin gas). Based on studies of deep-basin gas formation conditions, the theory of force balance and the equation of material balance, the favorable areas and gas-containing ranges, as well as possible gas-rich regions are preliminarily predicted in the deep-basin gas pools in the Upper Paleozoic He-8 segment of the Ordos basin.

Genetic Types and Distribution of CO_2 Gases in the Huanghua Depression

CO2 gas is a nonhydrocarbon gas, with a high economic value and a broad prospect for application. In the Huanghua Depression, there exist many genetic types of CO2 gases, i.e. organic CO2, thermal metamorphic CO2 and crust-mantle mixed CO2. The distribution of different types of CO2 gases is controlled by different factors. Organic CO2 that occurs mainly around the oil-generating center is associated with hydrocarbon gases as a secondary product and commonly far away from large faults. Thermal metamorphic CO2 occurs mainly in areas where carbonate strata are developed and igneous activity is strong, and tends to accumulate near large faults. CO2 of such an origin is higher in concentration than organic CO2, but lower than crust-mantle mixed CO2. Crust-mantle mixed CO2 occurs mainly along large faults. Its distribution is limited, but its purity is the highest.

Genetic type and source of natural gas in the southern margin of Junggar Basin, NW China

Natural gas has been discovered in many anticlines in the southern margin of the Junggar Basin. However, the geochemical characteristics of natural gas in different anticlines haven’t been compared systematically, particularly, the type and source of natural gas discovered recently in Well Gaotan-1 at the Gaoquan anticline remain unclear. The gas composition characteristics and carbon and hydrogen isotope compositions in different anticlines were compared and sorted systematically to identify genetic types and source of the natural gas. The results show that most of the gas samples are wet gas, and a few are dry gas;the gas samples from the western and middle parts have relatively heavier carbon isotope composition and lighter hydrogen isotope composition, while the gas samples from the eastern part of southern basin have lighter carbon and hydrogen isotope compositions. The natural gas in the southern margin is thermogenic gas generated by freshwater-brackish water sedimentary organic matter, which can be divided into three types, coal-derived gas, mixed gas and oil-associated gas, in which coal-derived gas and mixed gas take dominance. The Jurassic coal measures is the main natural gas source rock in the southern margin, and the Permian lacustrine and the Upper Triassic lacustrine-limnetic facies source rocks are also important natural gas source rocks. The natural gas in the western part of the southern margin is derived from the Jurassic coal measures and the Permian lacustrine source rock, while the natural gas in the middle part of the southern margin is mainly derived from the Jurassic coal measures, partly from the Permian and/or the Upper Triassic source rocks, and the natural gas in the eastern part of the southern margin is originated from the Permian lacustrine source rock. The natural gas in the Qingshuihe oil and gas reservoir of Well Gaotan-1 is a mixture of coal-derived gas and oil-associated gas, of which the Jurassic and Permian source rocks contribute about half each.

Optimization of Fairhurst-Cook Model for 2-D Wing Cracks Using Ant Colony Optimization (ACO), Particle Swarm Intelligence (PSO), and Genetic Algorithm (GA)

The common failure mechanism for brittle rocks is known to be axial splitting which happens parallel to the direction of maximum compression. One of the mechanisms proposed for modelling of axial splitting is the sliding crack or so called, “wing crack” model. Fairhurst-Cook model explains this specific type of failure which starts by a pre-crack and finally breaks the rock by propagating 2-D cracks under uniaxial compression. In this paper, optimization of this model has been considered and the process has been done by a complete sensitivity analysis on the main parameters of the model and excluding the trends of their changes and also their limits and “peak points”. Later on this paper, three artificial intelligence algorithms including Particle Swarm Intelligence (PSO), Ant Colony Optimization (ACO) and genetic algorithm (GA) has been used and compared in order to achieve optimized sets of parameters resulting in near-maximum or near-minimum amounts of wedging forces creating a wing crack.

NOVEL APPROACH TO LOCATOR LAYOUT OPTIMIZATION BASED ON GENETIC ALGORITHM

Proper fixture design is crucial to obtain the better product quality according to the design specification during the workpiece fabrication. Locator layout planning is one of the most important tasks in the fixture design process. However, the design of a fixture relies heavily on the designerts expertise and experience up to now. Therefore, a new approach to loeator layout determination for workpieces with arbitrary complex surfaces is pro- posed for the first time. Firstly, based on the fuzzy judgment method, the proper locating reference and locator - numbers are determined with consideration of surface type, surface area and position tolerance. Secondly, the lo- cator positions are optimized by genetic algorithm（GA）. Finally, a typical example shows that the approach is su- perior to the experiential method and can improve positioning accuracy effectively.

Reconstruction of the linac photon spectrum based on prior knowledge and the genetic algorithm

In order to derive the linac photon spectrum accurately both the prior constrained model and the genetic algorithm GA are employed using the measured percentage depth dose PDD data and the Monte Carlo simulated monoenergetic PDDs where two steps are involved.First the spectrum is modeled as a prior analytical function with two parameters αand Ep optimized with the GA.Secondly the linac photon spectrum is modeled as a discretization constrained model optimized with the GA. The solved analytical function in the first step is used to generate initial solutions for the GA’s first run in this step.The method is applied to the Varian iX linear accelerator to derive the energy spectra of its 6 and 15 MV photon beams.The experimental results show that both the reconstructed spectrums and the derived PDDs with the proposed method are in good agreement with those calculated using the Monte Carlo simulation.

基于GA的RBF神经网络气液两相流持液率预测模型优化

为了提高气液两相流持液率预测精度,针对传统径向基函数(RBF)神经网络预测气液两相流持液率网络拓扑结构困难和收敛速度慢等问题,提出一种基于遗传算法(GA)优化径向基函数神经网络的气液两相流持液率预测模型。通过系统聚类算法和灰色关联度分析(GRA)对收集的实验数据进行处理,优选出最优模型特征,同时结合遗传算法确定了RBF神经网络结构参数。基于室内实验数据进行训练,并与常用于持液率预测的反向传播(BP)神经网络、GA-BP神经网络及RBF神经网络进行对比,评估了模型的准确性及可行性。结果表明:GA-RBF神经网络模型均方误差为0.0017,均方根误差为0.0416,平均绝对误差为0.0281,拟合度为0.9483。相较于其他神经网络模型,该预测模型表现出更高的计算精度和更强的泛化能力。

基于SPA-GA-SVR模型的土壤水分及温度预测

土壤湿度和温度是影响水文循环和气候变化的重要参数,在农业实践活动和生态平衡中起着重要作用。为及时、准确地监测土壤含水量(Soil Moisture Content,SMC)及温度,提出了一种基于高光谱数据的预测方法。实验数据集来自为期5天的实地测量,所获得的高光谱数据包含大量的噪声及冗余信息,因此首先用Savitzky-Golay卷积平滑对光谱数据进行降噪处理,利用连续投影算法(Successive Projection Algorithm,SPA)提取数据特征波长,然后通过遗传算法(Genetic Algorithm,GA)对支持向量机回归(Support Vector Regression,SVR)的超参数权值和偏置进行优化,构建SPA-GASVR混合算法模型对土壤水分和温度进行预测,并与BP神经网络(Back Propagation Neural Network,BPNN)、SPA-BP、SVR、SPA-SVR、GA-SVR这5种模型的预测性能进行比较。实验结果表明:各模型在土壤湿度低于30%的情况下,表现出的预测能力差异并不显著。但整体上,复合模型相比于单一的神经网络或机器学习模型具有明显的优势,且经过连续投影算法优化的模型进一步的提高其预测能力,最终SPA-GA-SVR算法在各项指标上均优于其他模型,土壤水分预测模型的R^(2)=0.981、RMSE=0.473%,土壤温度预测模型R^(2)=0.963、RMSE=0.883℃。实验证明基于高光谱数据,经过SPA和GA优化的SVR模型能实现对土壤湿度和温度精准的预测。该方法具有一定的应用价值和现实意义,可应用于便携式高光谱仪和无人机上,实现对土壤水分和温度的实时监测,为今后的播种及灌溉提供理论参考。

基于GA-LSTM自适应卡尔曼滤波的路面不平度识别

准确、快速地识别出车辆当前行驶的路面激励信息,是实现智能底盘控制进而保证车辆平顺性的关键。针对传统路面不平度识别算法准确率低、自适应性差等问题,提出了基于遗传算法(genetic algorithm,GA)优化长短期记忆神经网络(long short-term memory networks,LSTM)自适应卡尔曼滤波的路面不平度识别算法。基于2自由度车辆悬架模型,通过灰色关联法选择LSTM神经网络的特征输入变量,并采用GA优化LSTM神经网络的模型参数以准确识别路面等级,并据此实时更新卡尔曼滤波器算法中的噪声矩阵,实现了在复杂路况下对路面不平度的自适应识别。仿真和试验研究表明,所提出的基于GA-LSTM自适应卡尔曼滤波算法能够快速准确的识别路面不平度与路面等级,与传统卡尔曼滤波算法相比,相关系数、均方根误差和最大绝对误差分别提高3.11%、37.5%和51.2%,表明所提算法对复杂工况具有很好的自适应能力。

RESEARCH ON THE MINIMUM ZONE CYLINDRICITY EVALUATION BASED ON GENETIC ALGORITHMS

A genetic algorithm (GA)-based method is proposed to solve the nonlinearoptimization problem of minimum zone cylindricity evaluation. First, the background of the problemis introduced. Then the mathematical model and the fitness function are derived from themathematical definition of dimensioning and tolerancing principles. Thirdly with the least squaressolution as the initial values, the whole implementation process of the algorithm is realized inwhich some key techniques, for example, variables representing, population initializing and suchbasic operations as selection, crossover and mutation, are discussed in detail. Finally, examplesare quoted to verify the proposed algorithm. The computation results indicate that the GA-basedoptimization method performs well on cylindricity evaluation. The outstanding advantages concludehigh accuracy, high efficiency and capabilities of solving complicated nonlinear and large spaceproblems.

Improved genetic algorithm for nonlinear programming problems

An improved genetic algorithm（IGA） based on a novel selection strategy to handle nonlinear programming problems is proposed.Each individual in selection process is represented as a three-dimensional feature vector which is composed of objective function value,the degree of constraints violations and the number of constraints violations.It is easy to distinguish excellent individuals from general individuals by using an individuals＇ feature vector.Additionally,a local search（LS） process is incorporated into selection operation so as to find feasible solutions located in the neighboring areas of some infeasible solutions.The combination of IGA and LS should offer the advantage of both the quality of solutions and diversity of solutions.Experimental results over a set of benchmark problems demonstrate that IGA has better performance than other algorithms.

A Hybrid Genetic Algorithm for the Traveling Salesman Problem with Pickup and Delivery

In this paper, a hybrid genetic algorithm （GA） is proposed for the traveling salesman problem （TSP） with pickup and delivery （TSPPD）. In our algorithm, a novel pheromone-based crossover operator is advanced that utilizes both local and global information to construct offspring. In addition, a local search procedure is integrated into the GA to accelerate convergence. The proposed GA has been tested on benchmark instances, and the computational results show that it gives better convergence than existing heuristics.

FORM ERROR EVALUATION OF CIRCLES BASED ON A FINELY-DESIGNED GENETIC ALGORITHM

A genetic algorithm(GA)-based new method is designed to evaluate thecircularity error of mechanical parts. The method uses the capability of nonlinear optimization ofGA to search for the optimal solution of circularity error. The finely-designed GA (FDGA)characterized dynamical bisexual recombination and Gaussian mutation. The mathematical model of thenonlinear problem is given. The implementation details in FDGA are described such as the crossoveror recombination mechanism which utilized a bisexual reproduction scheme and the elitist reservationmethod; and the adaptive mutation which used the Gaussian probability distribution to determine thevalues of the offspring produced by mutation mechanism. The examples are provided to verify thedesigned FDGA. The computation results indicate that the FDGA works very well in the field of formerror evaluation such as circularity evaluation.