Research Article
Conflict-Resilient Incremental Offloading of Deep Neural Networks to the Edge of Smart Environment
Algorithm 1
The CIODE Planning algorithm.
| (i) | Input DNN models, prediction files, network speed, available devices set (, ,..., ); | | (ii) | Output; | | 0 | function Planning: | | (1) | [], [], [], [], []; | | (2) | update ;//waiting queue of each device | | (3) | update ;//the execution time of each layer corresponds to each device | | (4) | create the neural network execution graph (); | | (5) | = shortest path (); | | (6) | device id; | | (7) | = ();//Sort according to the delay improvement | | (8) | update ; | | (9) | = makepair (, ); | | (10) | return; | | (11) | End; |
|