Automatic Teaching–Learning-Based Optimization: A Novel Clustering Method for Gene Functional Enrichments - Computational Intelligence Techniques for Comparative Genomics - page 22

Step 3: Elitist Teaching Phase — Update procedure amid with duplicate elimination
Evaluate the difference between current mean result and best mean result
according to Eq. ( 1 ) by utilizing the teaching factor T F
(a) If the new solution is better than the existing solution, then accept or
else keep the previous solution
(b) Select the solutions randomly and modify them by comparing with
each other
(c) Modify duplicate solution via mutation on randomly selected
dimensions of duplicate solutions before executing the next

generation

Step 4: Elitist Learners Phase
Update the learner

' s knowledge with the help of teacher ' s knowledge

according to Eq. ( 3 )

(a) If the new solution is better than the existing solution, then accept or
else keep the previous solution
(b) Replace worst solution with elite solution

Step 5: Elitist Learners Phase — Update procedure amid with duplicate elimination

Update the learner s knowledge by utilizing the knowledge of some other

learners according to Eqs. ( 4 ) and ( 5 ).

'

(a) Modify duplicate solution via mutation on randomly selected
dimensions of duplicate

solutions before executing the next

generation

Step 6: Stoppage Criterion
Repeat the procedure from Step 2 to Step 5 till the termination criterion is
met.

(a) If termination criterion is ful lled, then we get the nal value of the

solution or else repeat from Step 2 to Step 5.

4 Automatic Clustering Using Elitist TLBO (AutoTLBO)

The proposed automatic clustering using Elitist TLBO algorithm (AutoTLBO)
follows a novel integrated approach by assimilation of elitism and cluster evaluation
implanted into TLBO algorithm. Elitism is a mechanism used in this algorithm to
preserve the best individuals from generation to generation. By this way, the
algorithm never loses the best individuals found during the optimization process. In
this algorithm, replacing the worst solutions with elite solutions is done at the end
of learner phase. In the present work, duplicate solutions are modi ed by mutation

on randomly selected dimensions of the duplicate solutions before executing the