Section: New Results

Scheduling and Game Theory

A stochastic model of failures has been used to optimize the scheduling of checkpoints on desktop grids [28] .

We have also shown that non-cooperative scheduling can be considered harmful in collaborative volunteer computing environments [33]

Optimal scheduling and route selection have been investigated using a novel approach based on Lagrangian optimization. This result is inspired from flow control in multi-path networks and was used for multiple mag-of-tasks application scheduling on grids [61] .

In the similar context of broker-based networks of non-observable parallel queues, we provide lower bounds on the minimum response time. We introduce the “Price of Forgetting” (PoF), the ratio between the minimum response times achieved by a probabilistic broker and a broker with memory, that is shown to be unbounded or arbitrarily close to one depending on the coefficient of variation of the service time distributions. We also put our results in the context of game theory revisiting the “Price of Anarchy” (PoA) of parallel queues: It can be decomposed into the product of the PoA achieved by a probabilistic broker (already well understood) and the PoF [10] .