Optimized Task Offloading Policy in Given Sequence in Mobile Edge Computing

In the era of big data, tasks with complex computation emerge in large numbers. It is hard to solve them properly when the tasks are in a given sequence. So we should find a way to utilize the local CPU and the edge server on the premise of keeping the task order. We consider a task sequence in a user device and a base station that can receive offloaded task to solve. We find out the optimized offloading policy by cutting the task sequence directed acyclic graph (DAG) into basic units. The power and frequency allocated to every task can be also gained at the same time. Then we can perfectly know how to deal with the task chain. We choose to solve the original problem by decomposing it into sub-problems according with different scenarios. We show that our algorithm achieves the least consumed time among the compared algorithms. Simulations are performed to verify the proposed algorithm.