Search

This paper proposes a production model for quality surveillance, which is based on a cusum chart using the decision integral criterion. An algorithm for determining the economically optimum control parameters is presented. Some criticisms of the conventional ad hoc design of a cusum chart are made, and it is suggested that the average run lengths (ARL) at the rejectable quality level should be less than 2 when this model is applicable. A simplified version of the algorithm is also devised which 8gives control plans which are close to optimum. This scheme is suitable for application at the factory level. Also included is a table of ARL for some neglected, but useful, values of the decision interval and the process deviate. [ABSTRACT FROM AUTHOR]

This paper compares the accuracy, speed and storage requirements of Algorithm AS 60, published in this issue, with those of the algorithm given by Ortega (1967). [ABSTRACT FROM AUTHOR]

SUMMARY In an earlier paper, Garside (1965) gave an algorithm for calculating all subsets in multiple-regression analysis, thereby obtaining the subset of a given size that minimized the minimum residual sum of squares. This paper develops further the facilities in the earlier algorithm and introduces a new algorithm which, although slightly restrictive, is very much faster in execution. [ABSTRACT FROM AUTHOR]

This article provides information on a solution for comparative static problems in which qualitative information is available but quantitative information is not. An additional property of the technique set out in the study can be incorporated directly into the system. The solution technique has been organized in the form of an algorithm, devised for maximum efficiency in use. The solution algorithm is designed to determine the sign pattern or patterns of the solution of a system of linear equations Ax = 0 where the signs of the elements of the matrix A are known but the numerical values are unspecified. The algorithm follows definite steps. The algorithm can be used to examine whether a determinant is invariant in sign with respect to changes in the numerical values of its elements.

This paper describes the result of carrying out such a maximisation of a social welfare function prescribed for a model of the United Kingdom economy over the period 1957-66. The choice of social welfare function was a crucial part of the exercise and is discussed in some detail. Optimisation of the welfare function was achieved using modern control theory.[1] The next section of this paper describes the model; its dynamic properties are then evaluated in a subsequent section. Next, the optimisation goal is presented in terms of the model, and a social welfare function is chosen. The optimisation procedure is then applied and the results obtained are presented and discussed. Finally, consideration is given to the ways in which the procedures might in later research be modified as a result of the lessons learnt from this experiment. The results presented in this paper have shown that this model gives a good picture of how controls influence the economic system, despite the restrictions placed upon its size. The major cycles of the fitting period were reproduced and the model's poor ability at forecasting was a direct result of the primitive assumption of a fixed exchange rate. One advantage of its structure was that, apart from the control variables, it had only two exogenous variables. Although there was no quantitative way of assessing the dynamic properties of the model, the results of the dynamic multiplier tests were qualitatively satisfactory. On the other hand the attempt to use a control algorithm in order to suggest a policy which might have given better economic results at the end of 1961 must be admitted to have been a failure. The procedure correctly reflected the instructions given to it in the form of the social welfare function which was selected, but this function treated public and consumers' expenditure as perfect substitutes and omitted to take account of the desirability of certain items and overstated the undesirability of certain oth... [ABSTRACT FROM AUTHOR]

The article discusses the theory of classification and classification algorithms. These included a consideration of the synergistic value of the simultaneous use of cluster, network, and multidimensional scaling analysis by F. James Rohlf. The metric considerations in the k-means method of cluster analysis has been described. Laszol Orloci described information-type measures of divergence and their application in biological classifications. Richard B. McCammon discussed the dendrograph as a graphical aid to classification.

SUMMARY Maximum-likelihood estimation problems can be solved numerically using function minimization algorithms, but the amount of computing required and the accuracy of the results depend on the way the algorithms are used. Attention to the analytical properties of the model, to the relationship between the model and the data, and to descriptive properties of the data can greatly simplify the problem, sometimes providing a method of solution on a desk calculator. This paper describes how parameter transformation, sequential minimization and nested minimization can be used to solve particular problems. Applications to well-known problems of distribution fitting, quantal responses and least-squares curve fitting are described. The implications for computer programming are discussed. [ABSTRACT FROM AUTHOR]

The article focuses on the importance of computer information retrieval to the working scientists. Several scientists are responsible of gathering and maintaining references, citations and reprints in different forms, as well as in searching and indexing. Information retrieval will help them in saving time and effort because it can be easily used in searching and pointing important items and citations. The process only requires logical decisions and card input which provides limitless categorization possibilities.

The principle of minimum discrimination information estimation and associated techniques are used to consider the partitioning of a three-way contingency table when second-order interaction is present. Within the partitioned model a sub-hypothesis of logit linearity is also studied. Computational procedures utilize a generalized iterative scaling algorithm of Darroch and Ratcliff. [ABSTRACT FROM AUTHOR]

Presents an algorithm that performs the essential calculations of euclidean cluster analysis. Matrix of observations; Number of clusters; Initial cluster centers.

Proposes two minor simplifications of algorithm AS 37. Importance of the algorithm to regression computation; Features of the algorithm.