DASE7034 Chap.5 Sensitivity Analysis and Shadow Prices
Sensitivity Analysis and Shadow Prices
Sensitivity analysis describes local changes around an optimal basis. Objective-coefficient ranges preserve the current basic feasible solution, RHS ranges preserve the current basis and support shadow-price updates, and reduced costs describe the improvement needed for a non-basic variable to enter.
Outside an allowable range, re-optimisation is required.
Ask what is changing before using a report
Sensitivity outputs answer different local questions. An objective-coefficient range concerns the value or cost attached to one activity and tells when the current basis remains optimal. A right-hand-side range concerns the availability or requirement in one constraint and tells when the same basis remains feasible.
A reduced cost concerns a currently non-basic variable, while a shadow price concerns a constraint. Mixing these objects can produce a numerically tidy but conceptually invalid update.
Interpret a shadow price with its range
Within the allowable right-hand-side interval, the shadow price gives the marginal objective change per additional unit of the named resource, holding the remaining model structure fixed.
Multiply it by a proposed change only when the entire change stays inside the reported range and the units agree. A zero shadow price often indicates slack under the current solution, but it does not prove the resource will never matter.
Once the basis changes, another marginal value may apply and the model should be resolved.
Use reduced cost as an entry threshold
For a non-basic activity at zero, the reduced cost measures how much its objective coefficient must improve before it becomes attractive under the current basis, with sign interpreted according to the solver convention.
It is not the direct profit from introducing one unit, because feasibility adjustments in basic activities are already incorporated. For a basic variable, a reported reduced cost is normally zero. Translate the number back into the original units and direction before advising a price or cost change.
Respect locality and simultaneous changes
Allowable ranges are usually stated for one change at a time.
Several coefficient or resource changes can invalidate the simple interval even when each appears acceptable separately; a combined-change rule may offer a conservative check, but exact analysis requires re-optimisation. Structural changes such as a new variable, a new constraint or altered coefficients are not covered by a basic right-hand-side update.
A good answer states the baseline, parameter changed, permitted interval, computed effect and the point at which a fresh solve becomes necessary.
What this chapter covers
- 01
Objective coefficient ranges
- 02
Constraint RHS ranges
- 03
Reduced costs
- 04
Shadow prices
- 05
Limits of local analysis
Worked example · free
Update an objective inside its range
- 2Verify that the change lies inside the allowable interval.
- 2Keep the current decision values fixed.
- 3Compute 0.6 times 400 as the objective change.
Key terms
- Allowable Range
- The interval over which a reported local sensitivity relationship remains valid for the current basis.
- Reduced Cost
- The coefficient improvement required for a non-basic variable to become attractive to the optimal basis.
- Shadow Price
- The local objective change per unit increase in a constraint RHS while its allowable range holds.
Sensitivity Analysis and Shadow Prices FAQ
Do quantities stay fixed after an RHS change?
Not necessarily. The basis can remain the same while the values of basic variables change with the RHS. The shadow price updates the objective locally but does not by itself provide the new quantity vector.
Can two allowable changes be applied separately and added?
Not automatically. Individual ranges generally assume other data remain fixed. Simultaneous changes can move the basis even when each isolated perturbation appears allowable, so use an appropriate combined-change rule or resolve.
Is a positive shadow price the price to pay for a resource?
It is the model's marginal objective gain per resource unit within range. Compare that gain with the actual acquisition cost in consistent units before deciding what the organisation should pay.
Exam move
For every sensitivity question, write a four-column note: changed item, range test, protected object and permitted shortcut. Then state the breakpoint beyond which a resolve is needed. This prevents reduced costs, coefficient changes and shadow prices from being mixed. Add the original parameter, proposed change, allowable interval and measurement unit before calculating anything.
For a right-hand-side change, multiply the shadow price only over the portion that remains within its range and distinguish objective benefit from the purchase price of extra capacity. For an objective-coefficient change, state whether the current basis and decision values remain optimal even though the objective value changes.
For a non-basic activity, translate reduced cost into the direction and amount of improvement required for entry under the solver's convention. Practise one invalid case involving simultaneous changes and explain why separate one-at-a-time ranges do not automatically combine. End every calculation with a boundary sentence: beyond the stated interval, re-optimisation is required.
Then translate the result into a decision such as a maximum justified capacity purchase, a price threshold or evidence that a dormant activity remains unattractive.
Working through Sensitivity Analysis and Shadow Prices in DASE7034? Sia is AskSia’s AI Mathematics tutor — ask any DASE7034 Sensitivity Analysis and Shadow Prices 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.