Harmony Search in Therapeutic Medical Physics - Music-Inspired Harmony Search Algorithm: Theory and Applications - page 192

Table 2. Comparison of Harmony Search and genetic algorithm optimization - average number
of iterations and average total simulation time

Patient Optimization Method Iterations Simulation Time (s)

1 Harmony Search

Genetic Algorithm
% Difference

116
403
347%

2515
7870
313%

2 Harmony Search

Genetic Algorithm
% Difference

48
254
529%

940
4388
467%

3 Harmony Search

Genetic Algorithm
% Difference

25
149
606%

474
2313
488%

4 Harmony Search

Genetic Algorithm
% Difference

33
178
536%

437
2267
518%

5 Harmony Search

Genetic Algorithm
% Difference

62
255
411%

210
1200
571%

6 Harmony Search

Genetic Algorithm
% Difference

150
473
315%

438
1036
236%

7 Harmony Search

Genetic Algorithm
% Difference

40
154
385%

935
3793
406%

8 Harmony Search

Genetic Algorithm
% Difference

13
33
254%

64
113
176%

9 Harmony Search

Genetic Algorithm
% Difference

40
130
325%

129
459
355%

Average % Difference 412% 392%

3.3.1 Comparison of Harmony Search and Genetic Algorithm Optimization
Harmony Search and genetic algorithm were used to optimize nine patients using the
constraints specified in Table 1. Each algorithm run was repeated five times for each
patient in order to improve statistical significance. During each simulation, the opti-
mizer was allowed to run completely, such that both algorithms would reach the same
end result. The results of the comparison are presented in Table 2.
It should be noted that the average number of iterations in Harmony Search did not
include the iterations used to construct the harmony memory (HM). Likewise, the av-
erage number of iterations in genetic algorithm did not include the iterations used to
generate the initial population. The results show that Harmony Search, in the best