Dynamic Service Migration in Mobile Edge Computing Based on Markov Decision Process

This is the simulation code for the paper S. Wang, R. Urgaonkar, M. Zafer, T. He, K. Chan, K. K. Leung, "Dynamic service migration in mobile edge computing based on Markov decision process," IEEE/ACM Transactions on Networking, vol. 27, no. 3, pp. 1272 – 1288, Jun. 2019. (arXiv link: https://arxiv.org/abs/1506.05261)

The code runs best on MATLAB. It also runs on GNU Octave but the plot of instantaneous cost may be shown with lower granularity.

To reproduce the random walk result (Fig. 6 of the paper), run mainRandomWalk.m.

To reproduce the result with real base station locations (Fig. 8 of the paper), run mainRealCellLocation.m.

There are certain parameters in mainRandomWalk.m and mainRealCellLocation.m that can be changed for different experiments. The main algorithms are implemented in algorithms.m which is called within mainRandomWalk.m and mainRealCellLocation.m.

Real user traces are obtained from http://crawdad.org/epfl/mobility/20090224/ and the base station locations are obtained from http://www.antennasearch.com/. These are saved in traceRealCellLocations.mat.