In the frame of the MSc thesis I created a simulation environment, which integrates the function of a tight glycemic control (TGC) method with the function of a continuous glucose monitoring system (CGM) and a alarm regime for the proper interventions (STAR protocol- Sensor alarm settings). In the environment we can simulate the different sensor properties like sensor errors and measurement noises. Besides the different types of sensor errors the sensor noises were in the focus of interest; first of all a white-Gaussian noise was chosen. By the means of Monte-Carlo simulations the effect of the noise model was analysed based on the number of interventions and time-spans in the blood-glucose target range. Thereafter I expanded the analysis on other noise types and sensor properties. Clarke -error grid method was used to evaluate how the created virtual sensor differs from real ones. Finally the possibilities of the combined application of the two methods were evaluated with statistical methods.