204 Beam-based Correction and Optimization for Accelerators
10
-5
10
0
f
min
Rosenbrock-4, =0.001, N=500
GD
N-M Simplex
Powell
RSimplex
RCDS
GP (N=300)
10
-2
10
0
f
min
Rosenbrock-4, =0.01, N=500
20
40
60
80
100
cases
10
0
10
1
f
min
Rosenbrock-4, =0.1, N=500
Figure 7.13 Comparison of optimization performances for selected algorithms using
the Rosenbrock function with 4 variables. The sorted final minima of 100 runs are
shown for the noise levels σ = 0.001, 0.01, and 0.1 (from top to bottom).
10
-5
10
0
f
min
Rosenbrock-4, =0.001, N=500
GD
N-M Simplex
Powell
RSimplex
RCDS
GP (N=300)
10
-2
10
0
f
min
Rosenbrock-4, =0.01, N=500
20
40
60
80
100
cases
10
0
10
1
f
min
Rosenbrock-4, =0.1, N=500
Figure 7.13 Comparison of optimization performances for selected algorithms using
the Rosenbrock function with 4 variables. The sorted final minima of 100 runs are
shown for the noise levels σ = 0.001, 0.01, and 0.1 (from top to bottom).
