Abstract
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.