A Single Objective GA and PSO for the Multimodal Palmprint Recognition System

Algorithm 1

Proposed multimodal system using GA.

	Step 1: we initialize the number of chromosomes K as 50 and number of iterations as 100.
	Step 2: a binary string is generated which is equal to the size of the feature set, where bit 1 represents that the feature is selected and bit 0 represents that the feature is deleted.
	Step 3: K numbers of subsets are generated. The subset consists of , respectively (Figure 5). Here,. C is the original feature set.
	Step 4: we convert the genotype to phenotype using direct coding.
	Step 5: here, the recognition rate is the fitness function. It is calculated by the NN classifier. We consider two feature vectors A and B and . The Euclidean distance measure is used to calculate the distance between two vectors.
	Step 6: among the K chromosomes, K/2 chromosomes are selected by the roulette wheel method for the second iteration. Then, a uniform crossover with a probability of 0.7 is applied for the selected offsprings. We calculate the fitness function using step 5. We apply the mutation rate of 0.1 for all chromosomes and calculate the best fitness value until the last epochs are reached.