The analysis of queuing system based on support vector machine

The premise to evaluate the performances of a queuing system is based on knowing the distributions of customer arrival or service time in advance. It is very important to identify probability distributions or estimate density functions fast and efficiently. Support vector machine (SVM) based on statistical learning theory has been used generally in machine learning because of its good generalization ability. By using SVM we can classify and identity some probability distributions appeared in queuing system and solve the density function regression problem through using support vector regression (SVR). Some other problems need to be solved are formulated in the end.