Abstract:
This thesis presents a technique for designing a Thai virtual keyboard. By using the genetic algorithm, 93 Thai characters were arranged. A population of solution keyboards was evolved. During the evolutions, each keyboard was evaluated by using a fitness function that contains 200,000 Thai characters. The evaluation was based on the cost function that is constructed from digraph (time taking for typing two consequent characters), and the position of each character. The best solution founded by genetic algorithm is faster than the standard Thai keyboard Ket-ma-nee and Pat-ta-chote layouts.