Time-varying arrival rates
Model changing demand over the day instead of hiding peaks inside one average arrival rate.
A clinic may average 10 arrivals per hour across a day while receiving 18 per hour around lunch and 5 per hour late in the afternoon. A model using one constant rate can miss the peak queue completely.
A simple approach is a piecewise schedule: one arrival rate for each hour or operating period. A nonhomogeneous Poisson process is a standard model when arrivals remain independent but the rate changes with time.
λ(t) = arrival rate at simulation time t
Hourly bins may be enough for staffing by shift. A transport model may need much finer intervals. Do not create a noisy minute-by-minute rate from too little data.
Keep weekdays, weekends, seasons, or known event days separate when they have genuinely different demand patterns.
Before trusting downstream waiting times, compare simulated arrivals by time period with the observed counts used to build the profile.