The problem of optimizing an active noise control system for the implementation of a multi-channel ANC system with the aim of global reduction of broadband noise in a telephone kiosk is addressed in this paper. This optimization involves finding best locations for loudspeakers and microphones, and optimizing control signals. The problem of finding a proper size for control system, i.e. the number of loudspeakers and microphones involved in the control system is also investigated. The mean of acoustic potential energy in the enclosure in a frequency range of 50 Hz to 300 Hz is selected as a measure for optimization. Several genetic algorithms are proposed and compared to find the global minimum of this performance index. In order to have a better performance in reaching the global minimum, the parameters of these genetic algorithms are tuned, and the best genetic algorithm is selected among them. Numerical simulations of the acoustical potential energy and also sound pressure at the height where the head of a person may be located, confirms the optimality of the locations proposed by the genetic algorithm. Besides, the robustness of the optimized control system with respect to eventual changes in the location of primary and secondary loudspeakers and also microphones is shown with several simulations.