A Load Balancing Approach Based on Multiprogramming GA
YU Yan-fang · Jisuanji fangzhen · 2008
In the load balancing problem, load schedule algorithm is pivotal. It plays an important role in the whole balancing system. The paper proposes an effective server end load balancing algorithm based on a novel Multiprogramming Genetic Algorithm (MGA). The proposed approach models the natural evolution to search the optimal solution and it is a novel computational model for the load balancing problem. The strategy of multiprogramming (executes the crossover operation or mutation operation many times) and choosing the best resultant individuals has largely improved the performance of MGA. Through this schedule algorithm, request response time is lessened, the utility of servers CPU is increased, and performance of the system of load-balancing is enhanced. The final simulation result suggests that this proposed algorithm is feasible, correct and valid.