In steady-state simulation the output data of the transient phase often cause a bias in the estimation of the steady-state results. A common method is to cut off this transient phase. This paper presents a new heuristic used to detect the warm-up period in steady-state simulation output. An evaluation procedure is used to compare the presented rule, called EVR, with the method MSER-m known as the most sensitive rule in detecting bias and most consistent rule in mitigating its effects. The rules are applied to the output generated by M/M/1 queuing process and the performance of the methods is tested at 4 different levels of utilizations. Various measures of goodness are used to assess the effectiveness, consistency and confidence of the methods.

Abstract
 1. INTRODUCTION
 2. Truncation Rules
  2.1 Marginal Standard Error Rules (MSER and MSER-5)
  2.2 Exponential Variation Rate Rule (EVR)
 3. Experimental Details
  3.1 Experimental Model
  3.2 Performance Measures
  3.3 Evaluation Procedure
 4. Results
  4.1 In case of ρ=0.3, 0.5, and 0.7
  4.2 In case of ρ=0.9
 5. Conclusions
 REFERENCES

• Hyung-Sool Oh(Kangwon National University)
• Kyoung-Jong Park(Gwangju University)