Skip to content
Tutorials Center
My learning

Topics

Lesson 37 of 38 · Projects and Practice

Practice Set: Networks and Simulation

Six problems on tandem lines, Jackson networks, mean value analysis and simulation checks, each with a runnable calculator so you can verify your solution.

This second practice set covers the network and simulation chapters. As before, solve each problem by hand or with a calculator first, then use the code to check. The code applies the formulas, so you can modify the numbers to create new exercises.

Network problems

Problem 1 (tandem line). Customers arrive at 4 per hour to a line of three M/M/1 stations with service rates 6, 5 and 8 per hour. Find the mean total time in the system and which station is the bottleneck.

Problem 2 (Jackson network). An open network has two nodes. External arrivals: 3 per hour to node 1 and none to node 2. From node 1, 50 percent go to node 2 and the rest leave. From node 2, 20 percent return to node 1 and the rest leave. Service rates: μ_1 = 8, μ_2 = 6. Find λ_1, λ_2, the utilizations and the mean time in the network.

Python Check problems 1 and 2

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Problem 3 (closed network). A system with a CPU (demand 0.04 s), a disk (demand 0.06 s) and no think time has N = 3 jobs. Use mean value analysis to find the throughput and mean response time with N = 1, 2, 3, and the throughput limit.

Problem 4 (saturation). Interactive users have think time 5 s. The demands are 0.05 s and 0.10 s. At what number of users does the system saturate?

Python Check problems 3 and 4

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Simulation problems

Problem 5 (validating a simulator). Simulate an M/M/1 queue with λ = 0.6 and μ = 1, using Lindley's recursion with 20 independent replications of 20,000 customers each. Compute a 95 percent confidence interval for Wq and check that the exact value ρ/(μ - λ) = 0.9 is inside it. For 20 replications use the t value 2.093 (19 degrees of freedom).

Problem 6 (what if). Repeat with service times uniform on [0, 2] (mean 1, variance 1/3, so cs² = 1/3) and compare with the P-K prediction ρ(1 + cs²)E[S]/(2(1 - ρ)).

Python Check problems 5 and 6 (simulation, so results vary a little)

Press Run (or Ctrl + Enter) to see the result here.

Program input · what you'd type while the program runs

Type the answers before you press Run, one per line. Each time the program reads a line (with input()), it gets the next line from this box.

Reflection

  • Which problem had a bottleneck, and how did you spot it?
  • In the simulation problems, what would make the confidence interval narrower: more customers per replication or more replications? (Both help; longer runs reduce bias, more replications tighten the interval estimate.)
  • Where would you expect decomposition and product-form formulas to fail?

Exact answers from formulas and simulated answers should agree within the confidence interval. When they do not, check units, the warm-up period, and whether the model assumptions (such as exponential service) match the code.

Check your understanding

Quick check · 4 questions

of 4 answered You scored / 4 · perfect!

  1. 1. In a tandem line with service rates 6, 5, 8 and λ = 4, which station has the highest utilization?

    Utilization is λ/μ, largest for the smallest μ: 4/5 = 0.8.
  2. 2. For the Jackson network, which arrival rate goes into Little's Law for the whole network?

    Only arrivals from outside count; internal transfers are not new customers.
  3. 3. A closed system with demands 0.04 and 0.06 s has a throughput limit of

    1/D_max = 1/0.06 ≈ 16.67.
  4. 4. To reduce the width of a confidence interval you can

    The half-width shrinks like 1/√R.

Summary

  • Tandem: the bottleneck has the smallest service rate; W = Σ 1/(μ_i - λ).
  • Jackson: solve the traffic equations, then each node is M/M/1; use external γ in Little's Law.
  • Closed networks: MVA recursion, throughput limit 1/D_max, saturation N* = (D + Z)/D_max.
  • Simulation: replicate, build a confidence interval, compare with theory.

Next: Project: A Call Center Staffing Tool