Journal of Operational Research in its Applications
Year:2010 | Volume:7 | Issue:3 (26)|Papers Count:10

In this paper, a VSS scheme with three adaptive sample sizes is proposed. The model is constructed based on the Markov chain approach and optimized using genetic algorithm optimization method. Finally, a comparison between the proposed VSS scheme and the one that is proposed by Faraz and Moghadam (2008) is made to investigate whether more improvements can be achieved by applying three adaptive sample sizes or two ones are adequate.

In this paper, we consider some cubic-order corrections of classic Newton’s method with Halley’s method and produce the combined methods by using Gauss’ numerical differentiation and numerical integration methods. Analysis of convergence verifying that the presented methods have five or six convergence order. Numerical tests satisfying these methods are more useful than the others.

How to close an inefficient DMU to the production possibility set frontier would be of functional and important value in DEA literature. To do this, some directions may be considered according to DM points of view, which could be known or unknown. This paper will attempt to present some DM points and also some algorithms to find the related directions. Two subjects will be discussed: input-output improvement by imposing the minimum changes on their values and investigating an appropriate direction regarding the priority importance between input and output factors.

One of the most interesting issues for data envelopment analysis researchers is the ranking of decision making units. So far, many approaches are introduced in this area that one of them is Sexton method. In this paper the mentioned method is modified so that the problem raised of alternative optimal solutions of the multiplier form has overcame. In other word, the obtained rank for units is unique. For demonstrate the reliability of the modified method and comparisons of the obtained results to AP ranking method some numerical example are given.

In this paper, using the linear form of FDH model, its dual programming problem is determined and the role of the non-Archimedean number ε is studied. Then, by solving n linear programming problem, an assurance interval is attached.

In 2002, H. Kotakemori, K. Harada, M. Morimoto and H. Niki are suggested a Gauss-Seidel method with the preconditioner (I+Smax), (see [8]). Based on their work, in this paper, a modified Jacobi method is introduced. The convergence rates of the new proposed preconditioned method is compared with Jacobi method by comparison theorem. In addition, numerical experiments are used to show the effectiveness of the new MGS methods.

In this paper, an approximate formula for calculating the inverse two-dimensional Laplace transformation on finite interval is provided. The variables u=1/p, u=1/q convert this transform into a Mellin convolution, with a transformed kernel involving the gamma function. The truncation of the infinite product representation of 1/Г(1-s)Г(1-t) leads to an approximate differential expression for the solution.

In this article the square of directed graph will be discussed. H2 is the square of graph H which with adding edges joins two separated vertexes with the distance of 2 (when H is a un directed graph). In this article the second power of directed graphs is studied for the first time and related theorem to this subject and some examples to explain it will be mentioned.