Network Dismantling on Signed Network by Evolutionary Deep Reinforcement Learning

. 2024 Dec 16;24(24):8026. doi: 10.3390/s24248026

Algorithm 5 DSEDR Algorithm

Input:
Initial population size: $n_{I}$ , maximum number of iterations $n_{g}$ , population size $n_{p}$ , parent solutions $P (q)$ , crossover probability $p_{c r o s s o v e r}$ , and mutation probability $p_{m u t a t i o n}$ .
1:
${P_{i} (0)}_{n_{I}} \leftarrow Initialisation (Algorithm 2) (n_{I})$ ;
2:
for q = 0 to $n_{g} - 1$ do
3:
$P {(q)}_{E A} \leftarrow Evolution (Algorithm 3) (n_{p}, P (q), p_{c r o s s o v e r}, p_{m u t a t i o n})$ ;
4:
$P {(q)}_{D R L} \leftarrow DRL (Algorithm 4) (P {(q)}_{E A})$ ;
5:
$V, Φ (V) \leftarrow MSS (Algorithm 1) (P (q) \cup P {(q)}_{E A} \cup P {(q)}_{D R L}, K, H)$
6:
Select population with the highest $Φ (V)$ and size of $n_{p} / 2$ from $P (q) \cup P {(q)}_{E A} \cup P {(q)}_{D R L}$ as $P {(q)}_{O p t}$ ;
7:
Randomly select population with the size of $n_{p} / 2$ from $P (q) \cup P {(q)}_{E A} \cup P {(q)}_{D R L} - P {(q)}_{O p t}$ as $P {(q)}_{R a n d o m}$ ;
8:
$P (q + 1) \leftarrow P {(q)}_{R a n d o m} \cup P {(q)}_{O p t}$ ;
9:
end for