Algorithm 2. Choice of a cooperative neighbour for relaying

\\ according to vi's view \\

1 vi initialises Q-values and π;

2 for each cooperative neighbour of vi, i.e., $v_j\in {N^{\prime }}_i$ do

3 | Q(vj) ← (1 − α2)Q(vj) + α2 · (vj);

4 end for

5 $\overline{B}\leftarrow \frac{1}{|{N^{\prime }}_i|}{\sum }_{v_j\in {N^{\prime }}_i}B(v_j)$;

6 for each cooperative neighbour of vi, i.e., $v_j\in {N^{\prime }}_i$ do

7 | π(vj) ← π (vj) + τ(Q(vj) − ¯);

8 end for

9 π ← Normalise(π);

10 vi selects a cooperative neighbour based on π;