THE PROBLEM OF CONFIDENCE AND THREE METHODS OF VARIANCE REDUCTION IN THE SIMULATION OF QUEUEING

SIMULATION MODELS HAVE BEEN WIDELY USED IN SOLVING PROBLEMS RELATED TO TRAFFIC FLOW THEORY YET, VERY OFTEN, THE RESULTS OBTAINED CANNOT BE RELIED UPON FOR ACCURACY. QUEUEING PROCESSES ARE COMMONLY SIMULATED BECAUSE OF THE DIFFICULTY OF FINDING SUITABLE MATHEMATICAL MODELS WHICH LEND THEMSELVES TO ANALYTICAL EXAMINATION. SUCH SIMULATIONS PRESENT PROBLEMS OF BIASED STARTING, SLOW CONVERGENCE AND INDETERMINATE CONFIDENCE IN THE RESULTS. THE PAPER REPORTS ON MEASURES OF EFFECTIVENESS IN REDUCING VARIANCE OBTAINED USING THREE METHODS OF ACCELERATING CONVERGENCE IN SIMULATIONS OF QUEUEING: (1) BY SAMPLING WITH RANDOM FRACTIONS AND THEIR ANTITHETICS; (2) BY REVERSING RANDOM FRACTIONS GENERATING ARRIVAL AND SERVICE DISTRIBUTIONS; AND (3) BY REGRESSION WITH A KNOWN PARALLEL PROCESS. THE PAPER EXAMINES THE CHANGES IN VARIANCE REDUCTION OBTAINABLE WITH DIFFERENT LEVELS OF UTILIZATION IN THE QUEUEING PROCESS. FINALLY A PRACTICAL METHOD IS PROPOSED AND TESTED, OF USING THE CENTRAL LIMITS THEOREM TO ESTIMATE CONFIDENCE INTERVALS. /AUTHOR/

• Supplemental Notes:
  • Vol 4, Part 1, PP 567-581
• Corporate Authors:

  ARRB

  Melbourne, Victoria  Australia 
• Authors:
  • Francki, M
• Conference:
  • Australian Road Research Board Conference Proc
• Publication Date: 0

Media Info

• Serial:
  • Issue Number: 0

Subject/Index Terms

• TRT Terms: Queuing; Randomization; Regression analysis; Traffic flow theory; Traffic simulation
• Subject Areas: Data and Information Technology; Highways; Operations and Traffic Management;

Filing Info

• Accession Number: 00227222
• Record Type: Publication
• Source Agency: Traffic Systems Reviews & Abstracts
• Files: TRIS
• Created Date: Jun 15 1970 12:00AM