258 Agent-Based Modeling and Simulation with Swarm
FIGURE A.1: PSO simulator.
A.2 PSO simulator and benchmark functions
The PSO search process could be observed using this simulator (se e
Fig. A.1). The user is able to freely define the functions for the search.
In addition, this simulator uses De Jong’s standard functions. De Jong’s
standard functions are benchmark tests for GAs, and are used for determining
the minimum value. The definitions of these functions, along with the defi-
nition fields and optimum values, are shown in Ta ble A.1. The form of the
functions and plots projected on the x
1
− x
2
plane are shown in Fig . A.2.
Benchmark functions F 4 and F 5 seem to be more difficult than the o thers.
The +GAUSS(0, 1) of F 4 shows the addition of values from the normal dis -
tribution w ith average 0 and dispersion 1. In other words, noise is included in
the various points in F 4 . With F 5, there is a series of 5 × 5 valleys lined up
in a g rid alignment, but the valleys do not have a uniform depth. The trough
of the valley at the lowermost left is the minimum value (≈ 1), while the local
minimum values of the remaining troughs increase sequentially from left to
right and from bottom to top, as 2, 3, etc. When leaving these tro ughs, it
rapidly approaches the maximum value of 500. Note that the coordinates of
a
ij
are as follows:
int a[2][25] = {
{-32, -16, 0, 16, 32, -32, -16, 0, 16, 32, -32, -16, 0, 16, 32,
-32, -16, 0, 16, 32, -32, -16, 0 16, 32},
{-32, -32, -32, -32, -32, -16, -16, -16, -16, -16, 0, 0, 0, 0,
0, 16, 16, 16, 1 6, 16, 32, 32, 32, 32, 32}
};
Originally, F 1, F 2, and F 3 can have generalized definitions with three or