Design of quantum dot semiconductor optical amplifier by intelligence methods

The outstanding features of Quantum Dot Semiconductor Optical Amplifiers (QD-SOA’s) such as all-optical signal processing and signal regeneration are caused them are widely used in the optical communication systems. Due to the nano structure of these amplifiers, accurate design and modelling of them is a complex and challenging problem. This paper addresses design a QD-SOA. We present a novel method based on genetic algorithm to design QD-SOA. Since it is essential to having a model for designing, a numerical model is obtained in the first step. Then, an artificial neural network model is made using training data sampled from numerical model. The experiments show the proposed neural model has high accuracy as well as low computation time. In the next step, we convert the design problem to a genetic algorithm problem. Using neural model, the suitable fitness function is defined. The user can optimize the cost of production by setting the weight for the design parameters. The proposed system finds the best solution that satisfies desired gain regarding the production cost