DASE7034 Chap.12 Queueing Variability and Jackson Networks
Queueing Variability and Jackson Networks
Open queueing networks combine external arrivals with routed customer flow. Traffic equations determine each node's effective arrival rate before node stability and waiting can be assessed.
Jackson-network decomposition relies on Poisson, exponential and routing assumptions, while variability approximations use coefficients of variation when arrival or service distributions are more general.
Solve traffic equations before queue formulas
External arrivals are only one part of a node's demand. Customers routed from other nodes, including feedback loops, also contribute to its effective arrival rate.
For each node, write effective arrivals as external arrivals plus the sum of upstream effective rates multiplied by their routing probabilities. Solve the simultaneous equations and check that every routing row, including the exit probability, sums to one.
Using external demand alone can substantially understate node utilisation and delay.
Test every node for stability
Once effective rates are known, compare each node's rate with its own total service capacity. A single unstable node prevents a finite steady-state network calculation even if aggregate capacity across all facilities appears sufficient. Compute node performance only after this check.
For a fixed route, expected times may be added; for branching and repeat visits, weight node exposure by visit probabilities or use throughput-based totals.
A network-wide Little's Law calculation provides a useful consistency check for the same entering population.
State what permits Jackson decomposition
An open Jackson network uses external Poisson arrivals, exponential service at nodes, probabilistic routing and the relevant independence and service assumptions.
Under those conditions, node queue lengths have a product-form steady state and each node can be analysed using its effective arrival rate. Blocking, finite buffers, shared servers, priorities or state-dependent routing can break the decomposition.
The method is powerful precisely because its assumptions permit a network problem to separate into node calculations.
Use variability evidence for intervention
When arrival or service processes are more general, squared coefficients of variation describe dispersion relative to the mean and can enter queueing approximations.
Appointment smoothing affects arrival variability, while standard work or triage may affect service variability. Capacity changes service margin, routing changes effective rates and variability changes waiting around the same averages. Compare these levers using consistent scenarios, recompute all affected traffic equations after a routing change and identify any node newly pushed near saturation.
What this chapter covers
- 01
External and routed arrivals
- 02
Traffic equations
- 03
Node stability
- 04
Jackson network decomposition
- 05
Arrival and service variability
Worked example · free
Solve a routed arrival rate
- 2Compute routed flow as one quarter of eight.
- 2Add the external stream of two.
- 1Report effective rate four per hour.
Key terms
- Effective Arrival Rate
- The total rate reaching a node from external demand and routing within the network.
- Jackson Network
- An open network of Markovian service nodes whose steady state has a product-form decomposition under stated conditions.
- Squared Coefficient of Variation
- Variance divided by the square of the mean, used to compare variability across time distributions.
Queueing Variability and Jackson Networks FAQ
Why are traffic equations solved simultaneously?
A node can receive customers from several other nodes, including feedback loops. Its outgoing throughput affects other arrivals, so effective rates are linked and generally must be balanced together.
Can node waiting times always be added?
Add them directly for a fixed serial route. In a branching network, weight each node's time by the probability or expected number of visits for the population being measured.
What can break Jackson decomposition?
Finite buffers, blocking, shared resources, priorities, non-exponential service or state-dependent routing can destroy the independent-node product form. Treat the Jackson result as a benchmark unless its assumptions match the system.
Exam move
Draw arrows with routing probabilities and write one traffic equation per node. Solve effective rates, test each node's capacity and only then compute waiting. Finish by identifying which node contributes most to total delay and whether routing makes that node a true bottleneck. Add an explicit exit probability to every routing row and verify that the row totals one.
Write one traffic equation per node with external arrivals separated from routed flows, solve the equations together and substitute the rates back to check balance. Compare each effective rate with that node's service capacity before calculating delay. For a branching route, compute visit probabilities or throughput weights instead of adding every node time to every customer's experience.
Cross-check total customers divided by total external arrival rate against the visit-weighted network time for the same entering population. Then vary one routing probability and recompute all affected rates, not only the destination node. In a second scenario, hold mean rates fixed while increasing arrival or service variability and explain which approximation term changes.
Conclude by naming whether capacity, routing or variability is the intervention supported by the evidence and which node becomes closest to saturation.
Working through Queueing Variability and Jackson Networks in DASE7034? Sia is AskSia’s AI Mathematics tutor — ask any DASE7034 Queueing Variability and Jackson Networks question and get a clear, step-by-step explanation grounded in how DASE7034 is taught and assessed. Read this chapter free, then take your hardest questions to Sia.