Shadow prices for modified least loaded routing (MLLR)

Author

Vargas, C. ; Hegde, M.V. ; Naraghi-Pour, M.

Author_Institution

Dept. of Electr. & Comput. Eng., Louisiana State Univ., Baton Rouge, LA, USA

fYear

1996

fDate

31 Mar-2 Apr 1996

Firstpage

45

Lastpage

49

Abstract

Presents modified least loaded routing (MLLR), a routing policy where no preference is given to the direct single link path but rather the first choice path is chosen statistically based on the results of an optimization problem. The authors calculate shadow prices for MLLR using the blocking probability obtained from a fixed point performance evaluation algorithm. Numerical results comparing performance of LLR and MLLR in, terms of blocking probability are introduced. As an application, the authors formulate and solve a nonlinear constrained optimization problem to calculate the load sharing coefficients that result in a maximum network rate of return for fixed levels of OD pair blocking probabilities

Keywords

optimisation; probability; stochastic processes; telecommunication network routing; telecommunication traffic; fixed point performance evaluation algorithm; load sharing coefficients; maximum network rate of return; modified least loaded routing; nonlinear constrained optimization problem; pair blocking probabilities; routing policy; shadow prices; Circuits; Constraint optimization; Exponential distribution; Intelligent networks; Maximum likelihood linear regression; Probability; Random variables; Routing; Switches;

fLanguage

English

Publisher

ieee

Conference_Titel

System Theory, 1996., Proceedings of the Twenty-Eighth Southeastern Symposium on

Conference_Location

Baton Rouge, LA

ISSN

0094-2898

Print_ISBN

0-8186-7352-4

Type

conf

DOI

10.1109/SSST.1996.493469

Filename

493469