An Ant Colony Algorithm Based Task Scheduling Strategy For Collaborative Execution Of Mobile Application Between Mobile Device And Cloud

Nowadays, mobile applications are gradually becoming heavy computational. Thus, there comes a new computing paradigm, MCC(Mobile Cloud Computing).A typical computing mode of MCC is cloud-assist computing by offloading. In this computing paradigm, how to schedule tasks between cloud and mobile devices is the most important problem.A common method to schedule mobile tasks is to transform the problem to shortest path searching problem under graph model and use a Lagrangian multipier to determine the task schedeling solution.This method can not guanantee to get the optimal scheduling solution. Sometime, the energy consumption of the scheduling solution produced by this method is not good. The state translation model proposed in this paper could well transform the task scheduling problem to the optimal state searching problem. As a good probabilistic artificial intelligent algorithm, ant colony algorithm is much suitable to solve this problem. A task scheduling algorithm based on ant colony is proposed and implemented. The algorithm continuously changes the “state” transferring probability to change the best state transferring probability. Thus by using this algorithm, the probability of getting the optimal task sechduling solution is much larger.By comparing to a typical mobile task scheduling algorithm, experiments are conducted to test the effectineness of the proposed ant colony based algorithm. The experimental results show that the proposed ant colony based algorithm is much more energy saving than the typical task scheduling algoritm with a 20-30% energy saving.