The active vibration control of flexible structures is a relevant topic in many applications, such as smart structures. Commonly, the placement of sensors and actuators is defined by some criterion and, then, the controller is designed to satisfy some specification. However, better results can be found when optimal placement of sensors and actuators is obtained by solving an optimization problem, considering, for example, the possible positions and the numbers of sensors and actuators available. In order to do that, in this paper, a discrete optimization technique, to determine the position of active elements is applied in an active control problem. Depending on the number of sensors and actuators and the candidate positions, it is impracticable to use a combinatorial algorithm because this is very expensive in terms of computational time due to the number of possible combinations. The technique presented in this paper is based on a "geometry cutting". The geometric cutting starts in the disturbance location and the number of active elements desired defines the number of regions. After this, in each new piece, a pair of sensor and actuator is allocated using a combinatorial optimization technique. The objective function used in this paper is the infinity norm that corresponds to the peak of vibration of the system. The example proposed is the sensors and actuators locations in a plate discretized using the finite element method. In this specific case, the objective of the algorithm is to obtain a solution of the locations (vertical or horizontal) of collocated sensors and actuators. All the computational implementation was done using the MATLAB software.