Resource allocation and adjustment with genetic algorithm in virtual computing systems

Resource allocation and adjustment models oriented load balance in virtual computing systems are studied. Resource allocation takes the minimum standard deviation of remaining resource utilization in physical machines as the goal to achieve load balance. Resource adjustment takes the load balance in physical machines and the cost of virtual machines migration as the goal. We construct a heuristic genetic algorithm to determine the solution of parameters in such two models. Experimental results show that (1) our models for resource allocation and adjustment are correct and meaningful in virtual computing systems; (2) our solution can effectively and rapidly solve the nonlinear objective functions with discrete values and multiple constraints; and (3) the results are able to satisfy the requirements of load balance in virtual machines migration. © 2011 ICIC International.