In the paper, the original method of balancing cell loads in wireless networks has been presented. It allows for dffusing the load to the area assembling cells with decreased load and, thus, improving the total network capacity. The method is highly effctive in load balancing and acceptable as far as computational complexity is concerned. Its performance has been proved by experimental tests.
2
Dostęp do pełnego tekstu na zewnętrznej witrynie WWW
A new approach based on Fourier analysis to study Diffusion-type methods for solving the load balancing problem is presented. This approach is applied on two newly introduced methods the Average Diffusion (ADF) and its extrapolated counterpart (EADF) for d-regular graphs. Necessary and sufficient conditions are found for their convergence, whereas the optimum value of the extrapolation parameter r is determined in terms of M and m, the second largest and smallest eigenvalues of the ADF iteration matrix. Instead of using matrix theory, we apply Fourier analysis to determine M and m in case the diffusion parameters Cij are not equal to a constant value, as is the usual assumption. Our approach is simpler than matrix analysis. Moreover, it is shown that EADF attains the same rate of convergence with the classical Diffusion (DF) method, exhibiting a better convergence behavior in certain applications (2D-torus) for large number of processors.
JavaScript jest wyłączony w Twojej przeglądarce internetowej. Włącz go, a następnie odśwież stronę, aby móc w pełni z niej korzystać.