Learning Simio Lab Series Lab Module 05 Entity Routing Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 1 Objectives and Outline • Lab Objectives – Continue learning basic Simio modeling – Learn two different methods to route entities in Simio • Link Selection Weights • Dynamic routing using node lists • Lab Outline – – – – Video 1 – Using Selection Weights for probabilistic routing Video 2 – Using Selection Weights for conditional routing Video 3 – Using Node Lists for dynamic routing Video 4 – Assignments Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 2 Video 1 – Using Selection Weights for Routing 20% Fail Arrive Adjust Inspect 80% Pass Depart • TV final adjustment and inspection process – – – – – TVs arrive at the rate of 20/hour (exponential interarrival times) Adjustment takes approximately 2 minutes (uniformly distributed between 1.75 and 2.25) Inspection takes approximately 1.75 minutes (exponentially distributed) 20% of inspected TVs are found to need re-adjustment Interested in Time In System, Number In System, Utilizations of Adjust and Inspect Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 3 Re-entrant Flow and Effective Arrival Rate p l 1-p 𝜆′ = 𝜆 + 𝑝𝜆′ ′ 𝜆 = 𝜆 1−𝑝 Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 4 Selection Weights w1 w2 wn Link selection probability for link i : 𝑝𝑖 = Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 𝑤𝑖 𝑛 𝑗=1 𝑤𝑗 5 Video 2 – Conditional Routing Using Selection Weights Maximum of 3 adjustments for a given TV 20% Fail Arrive Adjust Inspect 80% Pass Depart • TV final adjustment and inspection process – – – – – TVs arrive at the rate of 20/hour (exponential interarrival times) Adjustment takes approximately 2 minutes (uniformly distributed between 1.75 and 2.25) Inspection takes approximately 1.75 minutes (exponentially distributed) 20% of inspected TVs are found to need re-adjustment Interested in Time In System, Number In System, Utilizations of Adjust and Inspect Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 6 Video 3 – Dynamic Routing Using Node Lists Approx. 10 minutes Server1 Arrival rate: 60/hour Arrive Approx. 5 minutes Server2 Server3 Depart Approx. 1 minute Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 7 Video 4 - Assignment • Using the TV Adjust/Inspect model: – “Fix” the issue where the TV’s are adjusted the 4th (and inspected) time even though we know that they will be rejected after inspection. Hint: Start by inserting a Basic Node in the path from the Inspect server back to the Adjust server. – Create a reference property for the maximum number of adjustments allowed and develop an experiment that compares the configurations with values 1, 2, 3, 4, 5, 100. Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 8 Video 4 - Assignment • Using the dynamic routing model (from the third video): – Develop an experiment with 25 replications using a run length of 500 hours with a 250 hour warmup and responses for the three server utilizations, the entity time in system (TIS) and entity number in system (NIS) – Using the base model, compare the 5 performance metrics using the following routing alternatives: 1. Probabilistic routing using the selection weights 6/78, 12/78, and 60/78 2. Preferred order with capacity 5 buffers at each server 3. Using AssociatedStationOverload with no buffers (capacity 0 buffers) 4. Using AssociatedStationOverload with capacity 5 buffers 5. Using the shortest queue length with infinite capacity buffers Copyright © 2013 - Jeffrey S. Smith and Simio LLC | All Rights Reserved 9