Research on Video Server Placement and Flux Plan Based on GA

Abstract

The paper is about a research on a problem of placing video servers and planning fluxes in an existing network topology, which must meet both the broadband limits and user's needs. For this problem, we build a bilevel mode and design two algorithms, in which the genetic algorithm is used to optimize the upper target and the SPFA minimum cost flow algorithm is used for lower target. Experimental results show that, in a complex network with large nodes and dense links, the proposed algorithm can gain a better server deployment solution in a relatively short period of time.