Genetic Algorithms in Search, Optimization and Machine Learning
Genetic Algorithms in Search, Optimization and Machine Learning
| Hi-index | 0.00 |
In the last decade, Active Noise Control (ANC) has become a very popular technique for controlling low-frequency noise. The increase in its popularity is a consequence of the rapid development in the fields of computers in general, and more specifically in digital signal processing boards. ANC systems are application specific and therefore they should be optimally designed for each application. Even though the physical background of the ANC systems is well known and understood, efficient tools for the optimization of the sensor and actuator configurations of the ANC system, based on classical optimization methods, do not exist. This is due to the nature of the problem that allows the calculation of the effect of the ANC system only when the sensor and actuator configurations are specified. An additional difficulty in this problem is that the sensor and the actuator configurations cannot be optimized independently, since the effect of the ANC system directly depends on the combined sensor and actuator configuration. For the solution of this problem several other optimization techniques were applied, such as simulated annealing for example. In this paper the successful application of a Genetic Algorithm, an optimization technique that belongs to the broad class of evolutionary algorithms, is presented. The results obtained from the application of the GA are very promising. The GA was able to identify various configurations that achieved a reduction of 6.3 dBs to 6.5 dBs, which corresponds to an actual reduction of 50% of the initial acoustic pressure.