Painting Using Genetic Algorithm with Aesthetic Evaluation of Visual Quality - Technologies and Applications of Artificial Intelligence - page 126

1. Randomly determine the background color of each chromosome.
2. Chromosomes serve as agents to paint during their life time.
3. Mutation and crossover are performed on the chromosomes.
Note that in EvoEco the fitness of images is evaluated by users. When the
generated image is good, users can terminate the system. Figure 1 shows the
EvoEco system and some evolving images.
Fig. 1. Screenshot of the EvoEco system. The left frame shows the population of 16
chromosomes. The middle frame shows the history of previous eight choices. The right
frame shows the control panel.
3 The Proposed Method
Genetic algorithm (GA) is a well-known EA and has shown its effectiveness on a
variety of optimization problems. The general principle of GA is to simulate the
mechanisms of natural evolution, such as selection, crossover, and mutation [6].
Based on Darwin's theory “Survival of the Fittest,” GA is believed to be capable
of evolving candidate solutions into better ones. To solve an optimization prob-
lem, GA represents candidate solutions as chromosomes, for which the encoding
scheme is essentially related to the problem to be solved. Compared to conven-
tional search methods, GA manipulates a set (population) of chromosomes in
the search process. The fitness function is used to evaluate the quality (fitness) of
candidate solutions (chromosomes). For a maximization problem, higher fitness
implies better solution quality.