摘要:

In the case, where a given linear optimization problem is soluble, the simplex procedure due to DANTZIG leads after a finite number of steps to a, certain solution (to an optimal base-point), but thereby the question concerning the unicity of the solution is not answered. The intention of this paper is to give necessary and sufficient conditions for the unicity of .the solution in such a manner, that the unicity can be stated from the last tableau of the procedure. The case of degeneration should not be excluded.

ISSN:0047-6277    

DOI:10.1080/02331887008801002

出版商:Akademie-Verlag    

年代:1970

数据来源: Taylor

摘要:

Linear programming problems with coefficients of the objective function, depending non linear from one parameter, may be solved using the theory of multiparametric linear programming (cf. Sokolová [3]). The condition under which this theory can be used is a special kind of dependence for each coefficient that consists practically in representing each coefficient as a linear combination with constant and variable factors with a known variation (e.g. whether, variation of different economic factors etc.). One gets subsets of the set on which the curve appearing in such a manner in the parameter space is defined by intersecting this curve with certain maximal polyhedrons, arising by computing the associated multiparametric problem. All elements of such a subset have either the same optimal solution or an unbounded objective function. At least one gives a computing procedure for the described problems.

ISSN:0047-6277    

DOI:10.1080/02331887008801003

出版商:Akademie-Verlag    

年代:1970

数据来源: Taylor

摘要:

As khown, the much usedx2goodness-of-fit test has the disadvantage that the division into classes thereby extensively follows according to the user's arbitrariness. For an optimal choice of classes Mann and Wald [10] recommended the use of power function of thex2test. However, the restriction to the case of equiprobable classes and the high numbers of classes obtained thereby do not satisfy. By a more careful numerical analysis one can show that a much lower number of classes leads only to unimportant power losses. In this way, a thumb rule results for the practical man: At the level of significance α = 0,05 choice 16 classes, independently of the alternative. Furthermore, the optimal class division is in general not the equiprobable one, it can be found by using of dynamic programming.

ISSN:0047-6277    

DOI:10.1080/02331887008801005

出版商:Akademie-Verlag    

年代:1970

数据来源: Taylor

摘要:

Known methods for predicting one variable are generalized to the case of predicting a vector of endogenous variables in the multidimensional regression model. By using the information of all equations we get better predictions (in the sense of a fixed optimalization criterion) than by predicting the vector components one by one.

ISSN:0047-6277    

DOI:10.1080/02331887008801006

出版商:Akademie-Verlag    

年代:1970

数据来源: Taylor

摘要:

D. B. Judin, E. G. Golstein: Lineare Optimerung I, in deutscher Spraehe herausgegeben von Prof. Dr. W. Dück Akademie-Verlag, Berlin 1968, VII, 528 S., 38 Abb., 31 Tabellen, )M 34,50.

ISSN:0047-6277    

DOI:10.1080/02331887008801007

出版商:Akademie-Verlag    

年代:1970

数据来源: Taylor