Rappresentazione e valutazione del resto nelle formule di integrazione numerica generalizzate

In previous articles we proved that the truncation errorE(f) of a quadrature formula over the interval [a,b] is always expressible in the form $$E(f) = \lambda _1 f^{(m + 1)} (\xi _1 ) - \lambda _2 f^{(m + 1)} (\xi _2 )$$ withf∈C ^m+1 [a,b], λ ₁ #x03BB; ₂ known positive constants, ξ₁, ξ_{1 2}∈]a,b[ generally unknown,m precision degree. Further we gave a similar representation, obviously more complicated, for the truncation error of integration formulas over multidimensional intervals. Developing logically the above mentioned researches, here we consider composite integration rules; as a matter of fact it is possible, not only to extend the preceding results, but also to find characteristic asymptotic properties of the truncation error representation, deeply related to the subdivision of the integration interval. Some numerical evaluations point out the applicability of the obtained estimations.     

Ricerca del minimo vincolato di una funzione dotata di derivate prima e seconda

The paper deals with a method for the constrained minimization of an-variable non-linear function. The method is quite fast to find the minimum also in its neighborhood and therefore it may be used for the solution of non-linear systems. The way to deal with the linear or non-linear constraints makes calculation time reasonable also in difficult problems. The method is essentially based on the local second-order approximation of the objective function (and so it is exact for quadratic functions) and on diagonalization of the associated quadratic form.      

Interpolazione ed approssimazione simultanea di un insieme di dati di funzione e derivate prima e seconda. Descrizione del programma midia

The problem of the least square fitting with fixed constraints is here solved in the general case of mixed data, i.e. functions and/or derivatives up to the second order. Given a set of linearly independent functions, the coefficients, the error matrix and the variance of the linear combination at any given point are determined. The abscissa and function (or derivatives) values can be reduced in any interval and the metric is consequently transformed. Making use of a sub-matrices method, the decrease of the weighted sum of the square of the residual for fits with higher number of functions is proved. The MIDIA code, which applies the obtained formulae, has been written in FORTRAN language for the IBM-7094 computer. As independent functions the most common polynomials (monomials, Legendre, Hermite, Chebychev etc.…) are used. The program fits successively higher degree polynomials to input data and at each step the weighted sum of the square of the residual is given, and the one which will occur at the next step is predicted. The tabulation gives the values of the linear combination, of its first and second derivatives and of their standard deviations at any mesh point. If the number of input data is too large for an unique interpolation, an automatic disposal for grouping them is provided.      

Approssimazione bilaterale A-stabile nella integrazione di equazioni differenzali ordinarie

Following a previous paper [1] we estabilish conditions for the bilateral approximation of the solution of ordinary differential equations usingA-stable predictor-corrector formulas. An extension to parametric formulas is carried out and numerical examples concerned with stiff equations are presented.

---

---

Dedicato al Professor S. Faedo in occasione del suo settantesimo compleanno     

Procedimento ricorsivo per la minimizzazione di neuroni nella realizzazione di una memoria strutturale

This work is part of a research program, the aim of which is, first of all, to specify a particular procedure in order to learn the characteristics through which one may better describe a set of messages; secondly to memorize, in the most economical manner, the information thus attained; and finally to code the data previously obtained in an optimal manner. The second part of the program leads us here to the search of a system of structural rules that will allow us to substitute, with remarkable economy, the memorization of the total set of data with that of a reduced set and of the structural rules obtained. Our problem stems from the study of a particular case, but is handled in general terms so that the algorithm given here may be applicable also to structural analyses of the type one needs when the information is given by means of boolean matrices (as it happens in fact in many problems connected with graphs, sequential circuits, minimization of boolean expressions, pattern recognition, etc.…). The structural rules are connected to particular regularities which are found in the set of data. The algorithm consists in the search of the minimum number of rules such that their logical disjunction implies a property defining the set of data.      

基于多智能体模型的动车组分布式预测控制

以高速铁路普遍采用的动力分散式动车组为研究对象.针对动车组由若干动力单元相互耦合组成的结构特点,以各动力单元为智能体.结合动车组牵引特性曲线和实际运行数据,应用减法聚类和模式分类算法建立各智能体多模型集.依据各智能体网络拓扑结构和相互耦合约束关系,建立动车组多智能体模型.针对各智能体的耦合约束,采用PID和GPC相结合的平稳起动切换控制策略和基于邻域优化的多智能体分布式协调控制算法,实现各智能体对给定速度的同步跟踪.基于CRH380A型动车组运行数据的仿真结果验证了本文方法的有效性.     

Generalizzazione del metodo di Laguerre

Sommario Nella seguente nota costruiamo una classe di metodi di ordine dispari comunque elevato per il calcolo delle radici di un polinomioP(x) generalizzando il metodo di Laguerre e, come per questo, dimostriamo la globale convergenza per radici reali. Il metodo è stato provato per polinomi con radici complesse e, al pari di Laguerre, risulta usualmente convergente.

---

The following paper concerns the construction of a class of methods of every high odd order for the evaluation of the roots of a polynomial. Such methods are a generalization of the Laguerre“s method and for them, as for the Laguerre“s method, we proof the convergence for real roots. The method has been proved for polynomials with complex roots and, as the Laguerre“s method, it results usually convergent.

---

---

Lavoro svolto nell“ambito della attività del G. N. I. M. e dell“Istituto per le applicazioni del Calcolo.     

Sull'ottimizzazione di un sistema idro-termo elettrico mediante utilizzazione del principio del massimo su calcolatore analogico

Sommario Il presente lavoro tratta nn'applicazione del principio del massimo all'ottimizzazione dell'esercizio di un sistema di potenza, comprendente una centrale idro-eletrica e una centrale termo-elettrica connesse fra loro dal lato rete. Dopo un breve richiamo della teoria dei processi ottimali, detta teoria viene applicata al sistema dinamico in questione di cui si esaminano in dettaglio gli aspetti connessi con il trattamento su calcolatore analogico. Il presente lavoro ha posto in luce, fra l'altro, alcune difficoltà dovute alla presenza di condizioni restrittive sul prodotto di variabili di stato e di controllo; tali difficoltà, connesse con la determinazione delle condizioni di giunzione sulle variabili aggiunte, al raggiungimento della frontiera, non sono menzionate dalla letteratura.

---

This paper deals with an application of the maximum principle to the optimization of a hydro-termal system. Plants (a hydro-and a thermo-electric plant) are feeding power into a net work. After a very short survey of the theory of optimal processes, the theory has been applied to the above mentioned dynamic system. Computational aspects of the problem, connected with its treatment on an analogue computer, are examined in detail. This research points out some difficulties due to the presence of constraints on the product of state and control variables: these difficulties, connected with the determination of jump conditions on the adjoint variables when reaching the boundary, are not mentioned in the literature.

---

---

C.C.R.-Euartom-ispra-C.E.T.I.S.     

I modelli matematici nella RO,con particolare riguardo ai problemi relativi alla distribuzione dei prodotti

Relazione introduttiva al tema di studio delle “Giornate di lavoro 1964? dell'AIRO, svoltesi all'Università di Pisa nei giorni 29-30-31 ottobre 1964. (Cfr. Quaderni RO, Atti della quarta riunione scientifica, 1964, ed. Boringhieri).     

Gödel语言类型系统

G(o)del语言是一种说明性逻辑程序设计语言.该语言基于一阶逻辑,引入了一个多态多类的类型系统和多种新的语言成分,支持抽象数据类型和模块化程序设计等技术,语言本身也具有很强的说明性语义.详细介绍了G(o)del语言的类型系统及其构造,对在逻辑程序设计语言中引入类型系统的作用进行了一些深入的分析和讨论.     

Guarini, Juvarra e Antonelli. Segni e simboli per Torino

Valentina Filemio reports on a 2008 exhibit in Turin, Italy, that focussed on architects Guarino Guarini, Filippo Juvarra and Alessandro Antonelli. Palazzo Bricherasio, Turin, Italy 28 June – 14 September 2008