Algorithm 1: MCE-GA

Inputs: Population size list $PopNum$, evolutionary stagnation threshold $\epsilon$, the iterations $T$, interval steps of migration $migFr$, mutation probability list $p_m$, crossover probability list $p_c$

Outputs: Optimal offloading policy $A$, the total overhead $C$

1: Randomly initialize the populations $Pop$ and Elite population $Elite\_pop$

2: Initialize the inputs

3: For i = $1\hspace{1em}to\hspace{1em}PopNum.size$ do

4: $Pop$ = $Pop\left[i\right]$;

5: Evaluate the fitness value of each individual in the $i\text{-}th$ $Pop$;

6: Update Elite population $Elite\_pop$;

7: While stopping criterion is not met do

8: For i = $1\hspace{1em}to\hspace{1em}PopNum.size$ do

9:  offspring = Select ($Pop\left[i\right]$);

10:  pop = Cross and Mutate (offspring);

11:  $Pop\left[i\right]$ = $Pop\left[i\right]$+ pop;

12:  evaluate the fitness value of each individual in $Pop\left[i\right]$;

13:  Select individuals to get a new generation of population;

14: End For

15: IF evolutionary algebra % $migFr==0$ do

16:  Carry out population migration;

17: Update Elite population $Elite\_pop$;

18: Return optimal offloading policy and the total overhead.

19: End