Task offloading is a promising technology to exploit the available computational resources in spatially distributed fog nodes efficiently in the era of fog computing. In this paper, we look for an online task offloading strategy to minimize the long-term cost, which factors in the latency, the energy consumption, and the switching cost. To this end, we formulate a stochastic programming problem and the expectations of the system parameters are allowed to change abruptly at unknown time instants. Meanwhile, we consider the fact that the queried nodes can only feed back the processing results after finishing the tasks. Then we put forth an effective bandit learning algorithm, i.e. the BLOT, to solve this challenging stochastic programming under the non-stationary bandit model. We also demonstrate that our proposed BLOT algorithm is asymptotically optimal in a non-stationary fog-enabled network. Numerical experiments further verify the superb performance of BLOT.
Mobile offloading is becoming critical to meet the ever-growing amount of wireless data traffic and computing tasks. In this work, Prof. Xiliang Luo’s research group proposed one effective online learning algorithm for task offloading in the case of unkown system parameters. Main contributions are:
- Novel analytic model: An online learning framework is introduced to capture the unknown parameters in the non-stationary environment.
- Online decision with delayed bandit feedbacks: An online task offloading algorithm BLOT based on the delayed bandit feedbacks is proposed to balance the exploration-exploitation dilemma.
- Performance guarantees: An upper-bound on the number of non-optimal decisions and zero approaching regret behavior are provided.
This work was recently published by IEEE Transactions on Parallel and Distributed Systems, which is the very top journal in high-performance computing