Improved parameter optimization method for measurement device independent protocol

The optimal selection of parameters in practical quantum key distribution can greatly improve the generation rate and maximum transmission distance system. Owing to high cost global search algorithm, local algorithm is widely used. However, there are two shortcomings algorithm. One that solution obtained not always solution, other effectiveness dependent on choice initial value. This paper uses Monte Carlo method prove whether function convex, also simulates analyzes projection each dimension parameter, contrast approach previous article. In order eliminate effect value, this proposes particle swarm optimization which integrates first step use find a valid parameter leads nonzero rate, second adopt as value derive solution. Then, algorithms used conduct simulation their results compared. show non-convex because it does satisfy definition convex function, however, since has only one non-zero stagnation point, LSA still obtain with an appropriate When relatively long, invalid difficult effective by random method. overcome shortcoming system at slightly increasing complexity