| Input: The data set to be solved; |
| Output: The optimal feature subset; |
| (1) | Set related parameters, including the population size, N, the maximal iteration times, T, and , and so on; t = 0; |
| (2) | Cluster all the features into K clusters by using the method in Section 3.2. |
| (3) | Randomly generate N integer ideas or individuals. |
| (4) | Evaluate the fitness of each idea by equation (6); |
| (5) | While t < T |
| (6) | Grouping all the N ideas into M clusters by the method in [42]; |
| (7) | Select the best idea from each cluster as the cluster center; % the phase of updating ideas % |
| (8) | For i = 1: N % from the first idea to the last one |
| (9) | If a random number rand() < , then |
| (10) | Randomly select a cluster and determine its cluster center; |
| (11) | If a random number rand() < , |
| (12) | Select the cluster center; |
| (13) | Generate a new idea by the equation (7); |
| (14) | Implement the proposed disturbance operator; |
| (15) | Else |
| (16) | Randomly select a normal idea from this cluster |
| (17) | Generate a new idea by the equation (7); |
| (18) | Implement the proposed disturbance operator; |
| (19) | End if |
| (20) | Else |
| (21) | Randomly select two clusters; |
| (22) | If a random number rand() < , then |
| (23) | Select two cluster centers; |
| (24) | Generate a new idea by the equation (7); |
| (25) | Implement the proposed disturbance operator; |
| (26) | Else |
| (27) | Randomly select two normal ideas from the two clusters respectively; |
| (28) | Generate a new idea by the equation (7); |
| (29) | Implement the proposed disturbance operator; |
| (30) | End if |
| (31) | End if % the phase of selecting elite ideas % |
| (32) | Evaluate the new idea and update correspond old idea by Section 3.3. |
| (33) | End for |
| (34) | End while |
