Advances in Mobile Cloud Computing Systems

Toward Energy-Efficient Task Execution in Mobile Cloud Computing 65

cloud execution, as a function of the application completion deadline T

d

.

If n is smaller than 3 (e.g., n =2), the cloud execution is more energy-

efficient when the delay deadline is below a threshold. However, if n is

larger than 3 (e.g., n = 3,4,5), the mobile execution is more energy-

efficient. Hence, under a fixed input data size L, the monomial order of

n can affect the optimal execution decision. Moreover, by optimally decid-

ing where to execute the application, a significant amount of energy can be

saved on the mobile devices. For example, for n =2, when the delay dead-

line is 10