A generalized computing paradigm based on artificial dynamic models for mathematical programming

In this paper a novel computing paradigm aimed at solving non linear systems of equations and finding feasible solutions and local optima to scalar and multi objective optimizations problems is conceptualized. The underlying principle is to formulate a generic programming problem by a proper set of ordinary differential equations, whose equilibrium points correspond to the problem solutions. Starting from the Lyapunov theory, we will demonstrate that this artificial dynamic system could be designed to be stable with an exponential asymptotic convergence to equilibrium points. This important feature allows the analyst to overcome some of the inherent limitations of the traditional iterative solution algorithms that can fail to converge due to the highly nonlinearities of the first-order conditions. Besides we will demonstrate as the proposed paradigm could be applied to solve non linear equations systems, scalar and multi-objective optimization problems. Extensive numerical studies aimed at assessing the effectiveness of the proposed computing paradigm are presented and discussed.     

A novel soft computing method for engine RUL prediction

Prognostics is an engineering discipline focused on predicting the Remaining Useful Life (RUL) of a system or a component using raw multimedia (sensor) data. This paper presents a novel machine learning model for this task, which includes a smart ensemble of gradient boosted trees (GBT) and feed-forward neural networks. It incorporates discussions on the poor performance of MLPs and the need of ensemble models. Initial stages of data exploration and pre-processing are also comprehensively documented. Experiments are performed on the four run-to-failure C-MAPSS datasets defined by the 2008 PHM Data Challenge Competition. It concludes by presenting evaluations of multiple prediction models like MLP, SVR, CNN & gradient boosted trees (GBT). Gradient Boosted Trees are efficient in the sense that they produce an encouraging scoring model with minimum effort and also return feature importance information. The proposed method uses stacking ensemble of feed-forward neural networks and gradient boosted trees, as first level learner, and, a single-hidden layer- fully-connected neural network as the meta learner. This ensemble provides better results than any of the models alone or weighted average of their predictions. The proposed method outperforms MLP, SVR, CNN and GBT.

     

A novel text and image encryption method based on chaos theory and DNA computing

In today’s world, the security of information is associated with valid and reliable encryption algorithms that we have used in our systems. Today, the latest methods for data encryption are based on DNA computing. In this paper, we consider a reliable data encryption algorithm (OTP) which is theoretically unbreakable, but it experiences some disadvantages in its algorithm. These drawbacks have prevented the common use of its scheme in modern cryptosystems. In this research, we include a logistic chaotic map as an input of OTP algorithm. So, the obtained result of ‘Matlab Simulation’ could prove the efficiency of proposed algorithm in image encryption. In addition to the cryptography of text files, we can propose an interesting encryption algorithm based on a chaotic selection between original message DNA strands and OTP DNA strands. Finally, the empirical results of our proposed algorithm will be compared with AES Open SSl algorithm.     

A separation method for DNA computing based on concentration control

A separation method for DNA computing based on concentration control is presented. The concentration control method was earlier developed and has enabled us to use DNA concentrations as input data and as filters to extract target DNA. We have also applied the method to the shortest path problems, and have shown the potential of concentration control to solve large-scale combinatorial optimization problems. However, it is still quite difficult to separate different DNA with the same length and to quantify individual DNA concentrations. To overcome these difficulties, we use DGGE and CDGE in this paper. We demonstrate that the proposed method enables us to separate different DNA with the same length efficiently, and we actually solve an instance of the shortest path problems. Masahito Yamamoto, Ph.D.: He is associate professor of information engineering at Hokkaido University. He received Ph.D. from the Graduate School of Engineering, Hokkaido University in 1996. His current research interests include DNA computing based the laboratory experiments. He is a member of Operations Research Society of Japan, Japanese Society for Artificial Intelligence, Information Processing Society of Japan etc. Atsushi Kameda, Ph.D.: He is the research staff of Japan Science and Technology Corporation, and has participated in research of DNA computing in Hokkaido University. He received his Ph.D. from Hokkaido University in 2001. For each degree he majored in molecular biology. His research theme is about the role of polyphosphate in the living body. As one of the researches relevant to it, he constructed the ATP regeneration system using two enzyme which makes polyphosphate the phosphagen. Nobuo Matsuura: He is a master course student of Division of Systems and Information Engineering of Hokkaido University. His research interests relate to DNA computing with concentration control for shortest path problems, as a means of solution of optimization problems with bimolecular. Toshikazu Shiba, Ph.D.: He is associate, professor of biochemical engineering at Hokkaido University. He received his Ph.D. from Osaka University in 1991. He majored in molecular genetics and biochemistry. His research has progressed from bacterial molecular biology (regulation of gene expression of bacterial cells) to tissue engineering (bone regeneration). Recently, he is very interested in molecular computation and trying to apply his biochemical idea to information technology. Yumi Kawazoe: She is a master course student of Division of Molecular Chemistry of Hokkaido University. Although her major is molecular biology, she is very interested in molecular computation and bioinformatics. Azuma Ohuchi, Ph.D.: He is professor of Information Engineering at the University of Hokkaido, Sapporo, Japan. He has been developing a new field of complex systems engineering, i.e., Harmonious Systems Engineering since 1995. He has published numerous papers on systems engineering, operations research, and computer science. In addition, he is currently supervising projects on DNA computing, multi-agents based artificial market systems, medical informatics, and autonomous flying objects. He was awarded “The 30th Anniversary Award for Excellent Papers” by the Information Processing Society of Japan. He is a member of Operations Research Society of Japan, Japanese Society for Artificial Intelligence, Information Processing Society of Japan, Japan Association for Medical Informatics, IEEE Computer Society, IEEE System, Man and Cybernetics Society etc. He received PhD from Hokkaido University in 1976.     

基于并行计算的大规模群体行为建模与仿真方法研究

为解决当前群体行为模型因规模扩大而导致计算量剧增的问题,采用并行离散事件方法构建了大规模群体行为模型,利用YH-SUPE仿真引擎实现了群体行为模型的并行计算。重点介绍了模型中仿真对象和仿真对象信息交互的设计方法,并对该模型在不同数量的节点和仿真实体的环境下进行了测试。实验结果表明,将并行计算引入群体行为建模之中,可以显著提高仿真个体的数量,更加有效地支持了群体模型的实时运行。     

A complex mathematical modeling method for biological objects. Modeling the tundra community

We consider a method for mathematical modeling of ecologo-biological systems based on computational studies that unites formal and informal, analytic and imitational approaches. The method is based on complex studies that include a complete set of operations, from filtering biological information to constructing a set of interrelated models, including simplified ones, that admit an analytic (parametric) study. This lets us overcome the disadvantages of purely imitational approaches: they are restricted by numerical experiments and often have huge models. The proposed approach has been used to analyze animal population fluctuations with the tundra community model “vegetation-lemmings-arctic foxes.” As a result of our studies, we formulate hypotheses on leading mechanisms that determine the fluctuations of tundra animal populations.     

基于数学形态学的SAR图像分割方法

图像分割是遥感图像处理中很重要的一步。因SAP图像通常带有较强的嗓声，用传统的边缘检测方法效果不理想。作者利用数学形态学开闭运算和混合滤波，可据目标的形状选用算法中的探针，取得了较好的滤波去噪和目标分割的效果。     

A parallel computing and mathematical method optimization of CNN network convolution

Training convolutional neural network (CNN) is a compute-intensive task in parallel tolerance has become a complete training is very important. There are two obstacles in the distributed memory computing environment to develop a scalable parallel CNN. It depends on a small volume across the two model parameters shown adjacent to the height data. The other presents the maximum overlapping parallel computing inter-process communication, large amounts of data over a communication channel to go. They will be transferred to the calculation. Replication by using threads on each compute node to initiate communication available after gradient is achieved. Reverse spread output data is generated in each stage of the model layer, and data communication may be performed in parallel with the calculation of other layers. To impact the replication study's efficiency and scalability, evaluated the model structure and optimization of various mathematical methods. When using the image VGG- net model training dataset, use 256 and 512, respectively, small batch size to achieve speedup computing nodes.     

使用新混合模糊聚类算法的模糊系统建模方法

为了进一步提高模糊系统建立模型的精度,提出一种新的模糊系统算法ANFIS-HC-QPSO：采用一种混合型模糊聚类算法来对模糊系统的输入空间进行划分,每一个聚类通过高斯函数的拟合产生一个隶属度函数,即完成ANFIS系统的前件参数--隶属度函数参数的初始识别,通过具有量子行为的粒子群算法QPSO与最小二乘法优化前件参数,直至达到停机条件,最终得到ANFIS的前件及后件参数,从而得到满意的模糊系统模型。实验表明,AN-FIS-HC-QPSO算法与传统算法相比,能在只需较少模糊规则的前提下就使模糊系统达到更高的精度。     

A novel intelligent denoising method of ecg signals based on wavelet adaptive threshold and mathematical morphology

Due to high-frequency noise and low-frequency noise in ECG signals will interfere with the accurate diagnosis of cardiovascular diseases. With the intrinsic mode function (IMF), which is the main component indicators of high-frequency noise and low-frequency noise, this paper proposes an intelligent denoising method of ECG signals based on wavelet adaptive threshold and mathematical morphology. Firstly, this method performs Complete Ensemble Empirical Mode Decomposition with Adaptive Noise (CEEMDAN) for signals containing noise, and adopts zero-crossing rate to identify IMFs containing high-frequency noise and low-frequency noise. Secondly, according to the discreteness and randomness of IMF containing high-frequency noise, a wavelet adaptive threshold mathematical model is constructed. In this model, with the signal-to-noise ratio (SNR) improvement as the threshold adjustment parameter, the wavelet threshold is modified by niche genetic algorithm, and the optimal solution is obtained after removing high-frequency noise by wavelet decomposition and reconstruction. The waveform of IMF containing low-frequency noise changes slowly and its amplitude is large and it is difficult to remove low-frequency noise. Therefore, mathematical morphology is used to remove low-frequency noise. Finally, the intelligent denoising method of ECG signals is designed by superimposing denoised IMFs. MIT-BIH experiments show that in the process of removing high-frequency noise and low-frequency noise, compared with other denoising methods, the percent root mean square difference (PRD) and SNR improvement of the method proposed in this paper are improved, and the denoising effect is significant, which can provide expert knowledge and decision-making guidance for related application fields.

     

A spatial warping method for freeform modeling based on a level-set method

Providing an intuitive and effective tool for freeform geometric modeling is important for product design. We introduce in this paper a level-set based spatial warping method for freeform modeling, allowing shape deformation to be initialed by rigid body transformations of volumetric tools. Intuitive user operations including imprinting, deformation and smoothing are developed to shield the user from the underlying geometric complexity. Unlike mesh-based spatial warping methods, the developed method represents a digital model by implicit distance field data and describes its change of geometry by the level-set method. This guarantees the generation of topologically correct triangular mesh models and circumvents the error-prone remeshing and mesh-repairing processes, thus preventing topological errors such as self-intersections. We present this method with algorithm details, numerical experiments and modeling examples.     

A novel approximate method of computing extended Nash equilibria

In this paper, we propose a new method of computing an approximate Nash equilibrium with additional features. Existing algorithms often fail to produce an exact solution for games involving more than 3 players. Similarly, existing algorithms do not permit additional constraints on the problem. The principle idea of this paper involves proposing a methodology for computing approximate solutions through evolutionary computation. To do so, we first provide formal definitions of these problems and their approximate versions. Following which, we present the details of our solution. One of the most important advantages of the proposed solution is flexibility, which provides solutions to problems related to Nash equilibrium extensions. The proposed idea is tested on several types of games that vary with difficulty and size. All test sets are generated based on the well-known Gamut program. Additional comparisons with classical algorithms are also performed. Results indicate that Differential Evolution is capable of obtaining satisfactory solutions to large random and covariant games. The results also demonstrate that there is a high probability that even large games, in which a set of strategies with a non-zero probability of being chosen are very small, have a solution. The computation time depends mainly on the problem size, and the original Nash equilibrium problem is unaffected by additional modifications.     

A surface modeling method based on the envelope template

In this paper, we present a novel surface modeling scheme based on an envelope template. A two-parameter family of interpolating surfaces is generated by repeated bicubic interpolation of the given data points, and then a solution to the envelope condition and the envelope of the family are constructed. The continuity conditions of two adjacent patches along the common boundary are derived by analyzing the geometric properties of the envelope patch. In order to facilitate surface modeling, an envelope template is constructed, which has many desirable advantages including simple structure, good local features and so on. G² or C² composite surfaces can be obtained utilizing the envelope template sweeping over the data points.     

A novel, low power biosensor for real time monitoring of creatinine and urea in peritoneal dialysis

Novel biosensors, based on immobilized creatininase, creatinase and urease are developed, using ISFETs with weak inversion at pH 6–8 and 37.0 °C. The ISFETs with circuitry, demonstrate a linear relationship of urea and creatinine at the range of 0–200 and 0–20 mM, respectively. Preliminary results show that biosensors operating in weak inversion mode can eliminate many of the disadvantages of ISFETs operating in the strong inversion region, providing a wide dynamic range output in nanoAmp. Such characteristics fit the analytical requirements for improving real-time monitoring in peritoneal dialysis (PD). Further work covers stability of ISFET sensors biased with CMOS circuits in the weak inversion mode working in the room temperature of 15–40 °C.