Deep-Reinforcement-Learning-Based Joint Energy Replenishment and Data Collection Scheme for WRSN

. 2024 Apr 9;24(8):2386. doi: 10.3390/s24082386

Algorithm 1 Solve TSP by using simulated annealing algorithm

Input:
The coordinates of CH nodes $l_{i}$
Output:
The sequence of CH nodes $l_{n e w}$
1:
Initialize maximum iteration temperature $K_{0} = 50, 000$ , the sequence of CH nodes $l_{i}$ , and annealing coefficient $q = 0.98$ .
2:
$K \leftarrow K_{0}$ .
3:
while $K > K_{0}$ do
4:
Swap the order of any two nodes to generate a new visiting sequence $l_{n e w}$ .
5:
Calculating $l_{n e w}$ , denoted as $d f$ .
6:
if $d f > 0$ then
7:
if $e x p (\frac{d f}{K_{0}}) > r a n d (0, 1)$ then
8:
$l_{f} \leftarrow l_{n e w}$
9:
end if
10:
else
11:
$K = K \times q$
12:
end if
13:
end while