Input Modelling and Output Analysis
Input modelling determines distributions for simulation inputs. Output analysis draws valid statistical conclusions.
Data Collection
Gather inter-arrival times, service times, demand. Representative, sufficient, error-free data.
Distribution Fitting
Exponential, normal, Poisson, uniform, triangular. MLE or method of moments for estimation.
Goodness-of-Fit
Chi-square, Kolmogorov-Smirnov, Anderson-Darling tests. P-values guide acceptance.
Terminating Simulations
Independent replications with different seeds. Confidence intervals across replications.
Steady-State
Delete warm-up period. Batch means or independent replications. Welch's method for warm-up length.
Verification and Validation
Verification: runs correctly. Validation: represents reality. Both essential for credible results.
Summary
Rigorous input modelling and output analysis ensure statistically valid simulation results.