Understanding Simio Modeling and Experiments

Aug 19, 2024

Simio Module Introduction - Video 2 Notes

Overview

  • Continuation from the first video where an initial Simio model of a two-station flow line was built.
  • Objectives for this video:
    • Specify object properties for the model.
    • Introduce Simio experiment concepts.

Model Components

  • Source Object: Creates entities.
  • Servers: Process entities.
  • Sync Object: Tabulates results and destroys entities.

Arrival and Service Rates

  • Customer arrival rate: 25 per hour.
  • Service times:
    • Server 1: 2 minutes on average.
    • Server 2: 1.71 minutes on average.
  • Reality Check:
    • Service rates:
      • Server 1: 30 per hour (60 min / 2 min).
      • Server 2: 35 per hour (60 min / 1.71 min).
    • Arrival rate (25) is less than service rates (30 and 35), indicating the system is nominally stable.

Distribution of Times

  • Arrival and service times are assumed to follow an exponential distribution:
    • Arrival time mean: 2.4 minutes.
    • Service time mean for Server 1: 2 minutes.
    • Service time mean for Server 2: 1.71 minutes.
  • Use of Simio Expressions in modeling:
    • Random Exponential: Expression random.exponential(mean) is utilized.

Setting Object Properties

  • Source Object (Source One):
    • Set arrival time property to mean of 2.4 minutes.
  • Server Object:
    • Set processing time property for Server 1 and Server 2 using the Expression Builder.
    • Server 1: random.exponential(2).
    • Server 2: random.exponential(1.71).

Running the Model

  • Updated parameters were applied to the model.
  • Results are similar to previous runs, maintaining system stability.
  • A verification method checks if around 600 entities flow through the system over 24 hours.
    • Expected: Arrival rate (25) * Time (24) = 600.

Simio Expressions Overview

  • Common distributions available in Simio:
    • Uniform, Triangular, Exponential, Normal, Log Normal.
  • Expression Builder: Assists in building expressions by providing tooltips and sublists for ease of use.

Simio Experiments

  • Transition from interactive modeling to simulation experiments:
    • Set the ending time to a run length of 240 hours.
    • Observations should be around 6000 for this duration.
  • Explanation of Predictable Randomness:
    • The model yields consistent results due to the random number generator starting at the same point each run unless altered.

Running Experiments

  • Created a new experiment in Simio with default settings (10 replications).
  • Results show statistical summaries:
    • Minimum, Maximum, Average of observations across replications.

Conclusion

  • Next video will focus on using experiments for interpreting results.