Algorithm 2 ZD-Based Dynamic Base-Station Selection

1:Input: candidate base stations $\mathcal{B}$, transmitter payoff vector ${\mathbf{U}}_T$, warden payoff vector ${\mathbf{U}}_W$, target base station $B_r$;   2:Output: feasible ZD strategy $\mathbf{s}=(s_1,s_2,\dots ,s_{\mathsf{\Lambda }})$;   3:Construct the outcome set $\left\{{\omega }_{ij}\right\}$ and identify ${\mathcal{K}}_r$ and ${\overline{\mathcal{K}}}_r$;   4:Check the feasibility conditions in  (58) and (59) for the current target base station;   5:If feasible, solve the constrained ZD parameter problem and construct the ZD probability vector $\mathbf{s}$;   6:Otherwise, examine the remaining candidate base stations and select a feasible target according to (61);   7:If no feasible target exists, apply the randomized backoff policy in (62);   8:Periodically re-evaluate the ZD feasibility conditions;   9:Construct the transition matrix $\mathbf{Q}$ under the warden monitoring rule $\mathbf{b}$; 10:Iterate the stationary distribution and compute ${\overline{U}}_T$ and ${\overline{U}}_W$ by (49); 11:return $\mathbf{s}$;