Maximal profit service task partition and distribution in computer grid

The paper considers grid computing systems in which the resource management systems (RMSs) can divide service tasks into execution blocks (EBs) and send these blocks to different resources. The service price is determined as a function of time elapsed till the service task completion according to a fixed tariff. The service time distribution depends on the assignment of the execution blocks to the resources as well as on the reliability of these resources. The cost of service depends on the resources used for its execution and is proportional to the time when the resources perform the execution blocks assigned to them. The optimal task partition and distribution should maximize the provider’s profit, which is equal to the difference between the expected service price and its cost for the provider. The paper suggests an algorithm for solving this optimization problem. The algorithm is based on the universal generating function technique and on the evolutionary optimization approach. Illustrative examples are presented.