Advanced Analytics and Artificial Intelligence Applications



7. Experimentation and results

value of 10.

Table 4.

21

Figure 8.

discard values that exceed the allowed limits.

Pseudocode of "Search for a food source in the neighborhood of i."

Bio-Inspired Hybrid Algorithm for Web Services Clustering

DOI: http://dx.doi.org/10.5772/intechopen.85200

1027, it is obtained with 200 iterations.

Average results of collections executing 100, 200, and 500 iterations.

Experiments were carried out with three service collections, which involved grouping 50, 647, and 1027 Web services, respectively. For the creation of the similarity matrices, eight different semantic similarity measures were calculated for each collection of Web services. The resulting similarity matrices were filtered to

The determination of the parameters used by the algorithm was performed by brute force, resulting in using 10 and the value of "phi" (φ) set to 0.8; the value of the limit that makes up each vector from the beginning was established with the

In order to characterize the behavior of the algorithm, 20 executions were made with 100, 200, and 500 iterations of the algorithm. For each of the executions, the best value of f(x) found by the bees was determined. Subsequently, a statistical analysis of the results found was carried out. Table 4 shows the average values of f(x) for each of the instances with 100, 200, and 500 iterations, respectively.

Based on the results shown in Table 4, it can be affirmed that for the 50 services instance, the best values are found with 500 iterations, while for the instance of 647 services, the best values are obtained with 100 iterations. Finally, for the instance of

Matrix size 100 iterations 200 iterations 500 iterations

50 0.4631 0.5011 0.5169 647 0.4457 0.4152 0.4228 1027 0.4414 0.4782 0.4542

f(x) f(x) f(x)

#### Figure 6.

ABC algorithm pseudocode.


#### Figure 7.

Pseudocode of the "Generation of N initial food sources."

d.Limit is a counter that indicates the number of iterations that the employed bee has being exploiting the current food source. The employed bee is obligated to abandon the food source when a g number of iterations is achieved (Figures 6–8).

Bio-Inspired Hybrid Algorithm for Web Services Clustering DOI: http://dx.doi.org/10.5772/intechopen.85200


#### Figure 8.

Pseudocode of "Search for a food source in the neighborhood of i."

## 7. Experimentation and results

Experiments were carried out with three service collections, which involved grouping 50, 647, and 1027 Web services, respectively. For the creation of the similarity matrices, eight different semantic similarity measures were calculated for each collection of Web services. The resulting similarity matrices were filtered to discard values that exceed the allowed limits.

The determination of the parameters used by the algorithm was performed by brute force, resulting in using 10 and the value of "phi" (φ) set to 0.8; the value of the limit that makes up each vector from the beginning was established with the value of 10.

In order to characterize the behavior of the algorithm, 20 executions were made with 100, 200, and 500 iterations of the algorithm. For each of the executions, the best value of f(x) found by the bees was determined. Subsequently, a statistical analysis of the results found was carried out. Table 4 shows the average values of f(x) for each of the instances with 100, 200, and 500 iterations, respectively.

Based on the results shown in Table 4, it can be affirmed that for the 50 services instance, the best values are found with 500 iterations, while for the instance of 647 services, the best values are obtained with 100 iterations. Finally, for the instance of 1027, it is obtained with 200 iterations.


Table 4.

Average results of collections executing 100, 200, and 500 iterations.

d.Limit is a counter that indicates the number of iterations that the employed bee has being exploiting the current food source. The employed bee is

(Figures 6–8).

Pseudocode of the "Generation of N initial food sources."

Advanced Analytics and Artificial Intelligence Applications

Figure 6.

Figure 7.

20

ABC algorithm pseudocode.

obligated to abandon the food source when a g number of iterations is achieved
