Site validation of GPSS and SLAM II
Abstract
Scope and Method of Study: This study examines the computer simulations languages GPSS and SLAM II. A site validation of the PC version of SLAM II is performed. Results from twelve different models, each run in GPSS and in SLAM II, are compared. An analysis is made of the methods of pseudorandom number generation used in the two languages. GPSS and SLAM II normally distributed random number streams are subjected to a Chi-square test to confirm normality. Findings and Conclusions: The differences in output between identical models run in GPSS, mainframe SLAM II and PC SLAM II could not be attributed to anything other than statistical variations caused by the use of different random number streams. Mainframe SLAM II used approximately four times as much CPU time to run models as did GPSS. The seed values used in setting the SLAM II IBM PC version pseudorandom number generator must be between -32768 and 32767.
Collections
- OSU Master's Report [734]