Mobile Phone Usage Detection by ANN Trained with a Metaheuristic Algorithm

. 2019 Jul 14;19(14):3110. doi: 10.3390/s19143110

Algorithm 1 Pseudocode of the EA.

1: Define objective function.
2: Initialize population of epicenters.
3: Randomly initialize x $_{i}$ .
4: Initialize v $_{i}$ and

ρ_{i}

.
5: Define

μ

and

λ

.
6: Obtain and rank the first fitness values.
7: Initialize x $_{i}^{b e s t}$ and x $_{b e s t}$ with the fitness values.
8: Define Sr, and the Max. it.
9: while (it. < Max. it.) do
10: Randomly calculate

ρ_{i}

.
11: v $_{p}$ ←

\sqrt[]{\frac{μ}{ρ_{i}}}

12: v $_{s}$ ←

\sqrt[]{\frac{λ + 2 μ}{ρ_{i}}}

13: if (∣x $_{i}^{b e s t} ∣$ ≤ Sr) then
14:

v

\leftarrow v

s
15: else
16:

v

\leftarrow v

p
17: end if
18: Randomly select positive or negative v $_{i}$ .
19:

x_{i}^{t} \leftarrow x_{i}^{t - 1} + v_{i}

20: if (rand >v $_{p}$ ) then
21:

x_{i}^{t} \leftarrow x_{b e s t} + E x p_{μ} (s)

22: end if
23: Constrain x $_{i}$ if needed.
24: Obtain the new fitness values.
25: Actualize x

_{b e s t}

and x

_{i}^{b e s t}

26: it.← it.++
27: end while
28: Postprocessing and visualization of results.