Combining data augmentation,EDAs and grammatical evolution for blood glucose forecasting

The ideal solution for diabetes mellitus type 1 patients is the generalization of artificial pancreas systems. Artificial pancreas will control blood glucose levels of diabetics, improving their quality of live. At the core of the system, an algorithm will forecast future glucose levels as a function of food ingestion and insulin bolus sizes. In previous works several evolutionary computation techniques has been proposed as modeling or identification techniques in this area. One of the main obstacles that researchers have found for training the models is the lack of significant amounts of data. As in many other fields in medicine, the collection of data from real patients is not an easy task, since it is necessary to control the environmental and patient conditions. In this paper, we propose three evolutionary algorithms that generate synthetic glucose time series using real data from a patient. This way, the models can be trained with an augmented data set. The synthetic time series are used to train grammatical evolution models that work together in an ensemble. Experimental results show that, in a scarce data context, grammatical evolution models can get more accurate and robust predictions using data augmentation. In particular we reduce the number of potentially dangerous predictions to 0 for a 30 min horizon, 2.5% for 60 min, 3.6% on 90 min and 5.5% for 2 h. The Ensemble approach presented in this paper showed excellent performance when compared to not only a classical approach such as ARIMA, but also with other grammatical evolution approaches. We tested our techniques with data from real patients.     

Investigating the Use of the Variogram for Lattice Data

The standard tool used in the analysis of geostatistical data is the variogram. When the variogram is applied to lattice data, most commonly the data associated with each region are assumed to have been observed and arbitrarily assigned at the center or centroid of the region. Distances between centroids are then used to develop the spatial covariance structure through the variogram function directly. This arbitrariness of assigning the data to the centroid causes concern because the spatial structure estimated by the variogram depends heavily on the distances between observations. This article investigates what happens to the estimation of the variogram when each lattice value is, in fact, placed randomly within its associated region. We examine the effect that this randomly placed location has on the empirical variogram, the fitted theoretical variogram, and testing for the existence of spatial correlation. Both a regular lattice and an irregular lattice are used for demonstration. In particular, county level summaries of standardized mortality rates for lung, pancreas, and stomach cancer are investigated to see how placing data points randomly throughout the county affects the estimation of the variogram.     

An estimate of the attraction domain with a specified exponential stability index in a wheeled robot control problem

The problem of synthesizing a law for the control of the plane motion of a wheeled robot is investigated. The rear wheels are the drive wheels and the front wheels are responsible for the turning of the platform. The aim of the control is to steer a target point to a specified trajectory and to stabilize the motion along it. The trajectory is assumed to be specified by a smooth curve. The actual curvature of the trajectory of the target point, which is related to the angle of rotation of the front wheels by a simple algebraic relation, is considered as the control. The control is subjected to bilateral constraints by virtue of the fact that the angle of rotation of the front wheels is limited. The attraction domain in the distance to trajectory - orientation space, is investigated for the proposed control law. Arrival at a trajectory with a specified exponential stability index is guaranteed in the case of initial conditions belonging to the given domain. An estimate of the attraction domain in the form of an ellipse is given.     

AR（p）模型中的缺失数据估计

运用EM算法,对含有缺失数据的AR（p）模型进行参数估计,通过最大似然准则就非左端缺失的情况进行插补.最后,用蒙特卡洛方法给出实验分析,表明如下结果：（i）误差与AR模型的阶数正相关,与缺失比例正相关;（ii）当AR模型的特征根模长相对较小时,误差与数据长度负相关,且误差被控制在了标准差的30％以内;（iii）当模长中等时,误差基本控制在1个标准差左右;（iv）当模长较大时,误差与数据长度正相关,而且误差也相对较大.     

Discrete dynamic programming with outcomes in random variable structures

Multiobjective approach is the common way of generalization single-criterion dynamic programming models. Another way is to consider partially ordered criteria structures. That approach is rather rare. The aim of the paper is to present such a model. Generalization of Bellman’s principle of optimality is employed to create a forward procedure to find the set of all maximal elements. As this set is usual large, the second problem under consideration is to find its subsets. To reduce the number of solutions presented to decision maker we propose to apply a family of narrowing relations. That approach is similar to scalarization in multiobjective programming. Ordered structures of random variables based on mean–variance, stochastic dominance and inverse stochastic dominance are considered. Numerical illustration is given at the end of the paper.     

Torsional oscillations in symmetrical structures

A two-time perturbation technique is used to study the lateral and torsional motions of a nonlinear symmetrical structure subject to a lateral sinusoidal ground motion. It is shown that when the ground acceleration frequency is about one-third of or three times the natural frequency of the lateral motion, the symmetrical structure is particularly susceptible to torsional oscillations even when the natural frequency of torsion is not close to the natural frequency of the lateral motion. The implication of this type of nonlinear coupling between the lateral and torsional motions in real structures subject to earthquakes is also discussed.     

Self-triggered filter design for a class of nonlinear stochastic systems with Markovian jumping parameters

This paper is concerned with the self-triggered filtering problem for a class of Markovian jumping nonlinear stochastic systems. The event-triggered mechanism (ETM) is employed between the sensor and the filter to reduce unnecessary measurement transmission. Governed by the ETM, the measurement is transmitted to the filter as long as a predefined condition is satisfied. The purpose of the addressed problem is to synthesize a filter such that the dynamics of the filtering error is bounded in probability (BIP). A sufficient condition is first given to ensure the boundedness in probability of the filtering error dynamics, and the characterization of the desired filter gains is then realized by means of the feasibility of certain matrix inequalities. Furthermore, a self-triggered mechanism is designed to guarantee the filtering error dynamics to be BSP with excluded Zeno phenomenon. In the end, numerical simulation is carried out to illustrate the usefulness of the proposed self-triggered filtering algorithm.     

Computational error analysis of periodic solutions for singular semilinear parabolic problems

By using the monotone method, a theoretical and computational method is given to find, to the degree of accuracy desired, approximate solutions of a class of singular semilinear parabolic problems. So that the error between the actual solution and its approximation is within a given error tolerance, the number of iterations is determined. Since each iterate is in terms of an infinite series, the number of terms to be retained in each iterate is determined so that its error from the exact iterate is within a given error tolerance. An improved rate of convergence is then given to show that it is possible to reduce the number of terms retained in each iterate. An algorithm is also described to obtain numerical solutions. For illustration of the computational methods developed, a numerical example is given.     

(<Emphasis Type="Italic">T</Emphasis>, <Emphasis Type="Italic">S</Emphasis><Subscript><Emphasis Type="Italic">i</Emphasis></Subscript>) Policy Inventory Model for Deteriorating Items with Time Proportional Demand

An EOQ model is reconsidered here in which the demand rate is changing linearly with time and the deterioration is assumed to be a constant fraction of the onhand inventory. The planning horizon is finite and known and the replenishment periods are assumed to be constant. The problem is to find the optimal number of replenishments, which are instantaneous. When there is no deterioration, the model developed is related to the corresponding model for nondeteriorating items. An example followed by sensitivity analysis is given to illustrate the derived results.     

Non-Convex Strong Duality Via Subdifferential

In this article, the strong duality is treated. It is shown that the strong duality is equivalent to the non-emptiness of the subdifferential of a sort map involving the constraint functions. It is also noted that this technique is useful to verify the Assumption S. Indeed, the linearity of a constraint function h is not required as usually seen in the literature. Moreover, it is shown that this condition is easer to verify in the applications. We apply this new principle to the bi-obstacle problem, to the elastic-plastic torsion problem and to the continuum model of transportation.     

连续区间数的中位算子及其灵敏度分析

研究决策信息以区间数形式给出的方案决策问题。考虑到区间数非均匀分布的特点,为避免区间极端值在区间数确定中产生较大误差,本文将中位数概念运用到区间数的确定上,提出OIP(Ordered Interval Point)有序中位算子。取单位区间单调函数(BUM函数)为二分之一所表示的值为权重,将区间数确定为一个实数,并研究算子单调性和有界性的初等运算性质。通过比较OIP算子与COWA算子对态度参数和区间长度的反应灵敏度,获得了在一定条件下OIP算子对态度参数反应更稳健,对区间长度反应更灵敏的结论。最后用算例证明该算子的可行性和有效性。     

Modelling of shear localization in elastic material by means of energy relaxation

A new approach to the problem of shear localization is proposed. It is based on the energy minimization principle associated with micro–structure developments and the micro–shearing of a rank–one laminate which is aligned to shear band. The thickness of the shear band represented by the volume fraction is assumed to tend to zero. The non–convex energy due to the formation of the shear band is solved by energy relaxation in order to ensure that the problem is well–posed. An application of the proposed formulation to isotropic linear elastic material is presented. The capability of the proposed energy relaxation is demonstrated through numerical simulation of a plane strain tension. The numerical results demonstrate that there is no mesh–dependence. (© 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)     

基于熵的群决策专家选择研究

群决策是研究多人如何做出统一的有效抉择.在群决策中专家决策水平直接反应到决策结论上,是影响事业的成败关键,因此专家的选择对于群决策而言是至关重要的.建立了评价群决策专家的指标集,从专家选择角度定义了理想专家,给出了专家选择的熵模型,并通过一个例子说明该模型的有效性.     

Ito过程在免费邮箱硬盘容量动态分配模型中的应用

分析了目前电子邮箱市场的情况,运用微分方程建立了邮箱用户增长模型,预测某邮箱未来几年内的用户数量及企业需要提供的硬盘容量。从减少成本的角度分析,运用Ito随机过程,对风险进行评估,建立硬盘容量动态分配模型,指导企业如何安排合适的硬盘容量,既能满足市场需要,又达到规避风险和降低成本的目的,提高市场竞争力。     

3PLS物流设施网络评价的模糊数据包络方法

3PLS物流设施布局直接关系到企业的战略和运作两个层面,对若干初始方案进行评价是保证物流设施布局方案有效性的前提.针对物流设施布局方案评价问题的同质性和模糊性特征,利用三角模糊数表示评价过程中的模糊不确定因素,并提出基于模糊数据包络方法的方案评价框架.通过对算例的评价应用,说明了方法的合理性和有效性.     

A probabilistic programming model for blending aggregates

The paper introduces the formulation of a probabilistic programming model to find the optimum mix proportion of aggregates to meet the specific grading requirement in order to minimize the cost which consists of the material cost and the expected penalty cost. The model is probabilistic since the gradation, which is the major parameter, is a random variable. A linear programming model is first formulated. Using the LP solution as initial value, a direct search technique is then employed to solve the problem. The model is expected to be applicable to any problem of aggregates blending. In this paper, however, the mixing aggregates of an asphalt mixing plant is exemplified to test the applicability of the model.     

A method for the approximate solution of quasi-static problems for hardening bodies

A method for the approximate solution of quasi-static problems for hardening elastoplastic bodies is proposed. The constitutive relation of the model is taken in the form of a variational inequality. An approximate solution of the initial problem is constructed in time steps and, by means of the finite element method, is reduced to the solution of a system of two variational inequalities in corresponding finite-dimensional space. It is shown that the solution of this system is equivalent to finding the saddle point of the corresponding quadratic functional. To find the saddle point, Udzawa's algorithm is used, by means of which the process of finding the velocity vector and stress tensor reduces to the successive calculation of these quantities: the velocity vector is determined from the variational inequality corresponding to the equilibrium equation, and the stress tensor is determined from the variational inequality corresponding to the constitutive relation. The latter inequality is reduced to a certain non-linear equation containing the operation of projection onto a closed convex set corresponding to the elastic strains of the medium. In turn, the solution of the non-linear equation is constructed using the method of successive approximations. To illustrate the use of the proposed method, the one-dimensional problem of the quasi-static deformation of a cylindrical tube under a load applied to its internal surface is considered.     

The elastostatic three-dimensional boundary element method: analytical integration for linear isoparametric triangular elements

In this paper an analytical integration scheme is described that is designed to reduce the errors resulting from the numerical evaluation of integrals with singular integrands. The analytical scheme can be applied to linear triangular elements for use in elastostatic problems and is particularly useful for predicting distortion, to high accuracy, close to surfaces. It is demonstrated that although the analytical scheme takes longer computationally than the usual quadrature approach it is quicker when element subdivision is required to achieve reasonable accuracy. Numerical tests are performed on a simple test problem to demonstrate the advantages of the analytical approach, which is shown to be orders of magnitude more accurate than standard quadrature techniques.     

The wavelet multiscale method for inversion of porosity in the fluid-saturated porous media

In this paper, the wavelet multiscale method is applied to the inversion of porosity in the fluid-saturated porous media. The inverse problem is decomposed to multiple scales with wavelet transform and hence the original inverse problem is re-formulated to be a set of sub-inverse problem corresponding to different scales and is solved successively according to the size of scale from the smallest to the largest. On each scale, regularization Gauss–Newton method is carried out, which is stable and fast, until the optimum solution of original inverse problem is found. The results of numerical simulations demonstrate that the method is a widely convergent optimization method and exhibits the advantages of conventional regularization Gauss–Newton method methods on computational efficiency and precision.