The words contained in this file might help you see if this file matches what you are looking for:

...Why statistical significance test suppose we have developed an ec algorithm a want to compare with another b both algorithms are stochastic how can be sure that is better than assume run and once get the results x y respectively if minimisation it because or just of randomness treat as random number generator its output follows some distribution depends on seed collect samples many times independently using different seeds carry out tests based collected parametric non do not variables follow normal paired unpaired t z wilcoxon rank sum signed one sample he used when population mean standard deviation null hypothesis reject support statistically under level note exact critical value at use rough p for two tailed lower upper...