ITLS6002 Cheatsheet · Supply Chain Planning and Design

0 · Exam Mapcourse data · Canvas

“You can bring a double-sided A4 sheet of paper full of whatever notes you want into the exam.” (ITLS6002 Canvas final exam information, Semester 2 2026)

All material is assessable; the guide identifies particular emphasis on aggregate planning and the first network-design block. Excel is required. For every model: define the decision, keep units/time bases visible, make constraints auditable, interpret the operating consequence, then state what change would reverse the choice.

1 · Method Routertopic maps · FAQ

2 · EOQ + Reorder TriggerCh 1 · final model

Symbols and calculation

lot sizeQ* = √(2DS/H)
orders/year = D/Q · average cycle stock = Q/2
TRC(Q) = DS/Q + HQ/2

timingInventory position IP = on-hand + scheduled receipts − backorders
certain-demand ROP = dL
with uncertainty: ROP = mean lead-time demand + safety stock

Assumptions: constant demand; instantaneous replenishment; no stockouts; constant order and holding costs. Constant purchase cost drops from the basic objective because it cannot change Q*. It returns for quantity discounts. Lead time changes ROP, not basic EOQ.

Exam Report the mathematical Q*, then compare neighbouring feasible multiples. Equality of order and holding cost need not survive a carton, capacity or discount constraint.

3 · Safety Stock + ReviewCh 2 · policy comparison

Protection interval is the exposure the policy must cover. Continuous review can react immediately and protects L. Periodic review waits until the next review and protects R+L. Draw the interval before choosing mean and deviation.

independent period demandExposure T: mean = μT · deviation = σ√T
SS = zσ√T

continuous reviewT=L · ROP = μL + zσ√L

periodic reviewT=R+L · order-up-to S = μ(R+L)+zσ√(R+L)
order now = S − current inventory position

Raising the target raises z and safety stock. Price the consequence as incremental stock × annual holding cost. A service recommendation names the metric, protection, added carrying cost and assumption about demand dependence.

Policy sentence: “Review continuously and release Q when IP reaches R,” or “At each review order S−IP.” Add the units and the service metric. This prevents a correct target from becoming the wrong operational action.

4 · Advanced InventoryCh 3 · choose structure

All-units quantity discounts

1. For each price bracket compute H=iC and its EOQ.
2. Retain a feasible EOQ; otherwise retain the bracket's lowest qualifying quantity.
3. Compare full annual cost for every feasible candidate.

TAC(Q) = DC + DS/Q + iCQ/2

Reject a Q that does not qualify for its assumed price. Purchase cost and price-based holding cost both change across brackets.

Single-period newsvendor

Use only for one selling opportunity with defined underage/overage consequences—not normal carry-over or routine backorders.

Cᵤ = selling price − purchase cost
Cₒ = purchase cost − salvage value
critical fractile = Cᵤ/(Cᵤ+Cₒ)
choose first demand Q with F(Q) ≥ fractile

Time-varying lot sizing

Iₜ = Iₜ₋₁ + Qₜ − Dₜ
Qₜ ≤ capacityₜ × yₜ; yₜ ∈ {0,1}

The balance preserves flow. Binary yₜ activates the fixed setup cost and blocks positive quantity when no setup is released. Audit every period, non-negativity and the final position.

Installation stock is one location’s own inventory. Echelon stock includes that location and the downstream inventory it supports. Circle what each control point counts before assigning demand and lead-time exposure; otherwise the same protection can be treated as unrelated buffers.

5 · Aggregate Planning + S&OPCh 4 · final emphasis

Translate medium-term demand into production, workforce, overtime, subcontracting, inventory and backlog decisions. The model supplies trade-offs; S&OP reconciles demand, operations and finance and approves one feasible plan.

no-backlog material balanceending inventoryₜ = beginning inventoryₜ + regularₜ + overtimeₜ + subcontractₜ − demandₜ
beginning inventoryₜ₊₁ = ending inventoryₜ

Build → audit → recommend

1. Separate forecast, capacities, costs and opening conditions from decisions.
2. Reconcile each period before calculating cost. Keep one sign convention for inventory/backlog.
3. Check regular/overtime/subcontract limits, non-negativity and final-balance requirement.
4. Recompute every cost category from displayed quantities; compare plans on the same horizon, opening position, service rules and cost basis.
5. Explain the operating pattern; test demand, capacity or cost sensitivity; state a reversal condition.

6 · Facility + Distribution DesignCh 5–6 · final emphasis

Fixed-charge facility model

minimise ΣᵢFᵢyᵢ + ΣᵢΣⱼcᵢⱼxᵢⱼ
serve each zone: Σᵢxᵢⱼ = dⱼ
open/capacity link: Σⱼxᵢⱼ ≤ Capᵢyᵢ
xᵢⱼ ≥ 0; yᵢ ∈ {0,1}

Fixed cost rewards consolidation; transport, handling and service may reward proximity. Verify each destination separately—a correct network total can hide one short zone and one oversupplied zone. Report opened sites, flows, utilisation and cost; test demand growth, cost change, capacity loss and service constraints.

Location estimate + channel choice

centre of gravity: x̄ = Σxⱼwⱼ/Σwⱼ · ȳ = Σyⱼwⱼ/Σwⱼ

This generates a candidate, not a decision: test discrete sites, road access, capacity, fixed cost, risk and response requirements. For distribution, compare inventory aggregation, facility expense, transport, response time, assortment, customer effort, returns and scalability.

7 · Drill 1 — EOQ + ROPCh 1 worked · 8 marks

Given: D=9,600 units/year; S=$75/order; H=$6/unit-year; L=3 weeks; 48 operating weeks/year.

Decision: order about 490 units whenever inventory position reaches 600 units. The first number is how much; the second is when. At unconstrained EOQ, annual ordering and holding costs should be approximately balanced.

Trap Do not use 600 as the replenishment quantity or 490 as the trigger. A longer certain L moves the trigger but not basic EOQ.

7b · Feasible-Quantity CheckCh 1 decision check

For D=9,600, S=$150, C=$40 and i=25%: H=$10 and unconstrained EOQ=536.66. With cartons of 60, compare 480, 540 and 600 on annual ordering + holding cost. 540 has the smallest displayed relevant total; keep ROP as a separate timing decision.

7c · EOQ Answer AuditCh 1 FAQ

If price varies, restore purchase cost. Name any broken basic-model assumption.

8 · Drill 2 — ProtectionCh 2 worked · 8 marks

Given: weekly μ=120, σ=24 units; L=2 weeks; review R=3 weeks; z=1.645.

Decision: use ROP≈296 under continuous review, or target S≈688 at the three-week review. Periodic order quantity is 688−IP, not 688 automatically.

Checks Units → exposure → mean → deviation → z → SS → trigger/target → carrying cost. If demand is seasonal or correlated, do not scale one-period σ automatically.

8b · Service-Language TrapCh 2 clinic

Report both the service gain and annual carrying-cost consequence. More protection is not free.

8c · Sensitivity NarrativeCh 2 FAQ

R↑ raises periodic mean and variability exposure; z↑ raises only protection; σ↑ raises safety stock; μ↑ raises expected exposure. Continuous ROP is unchanged by R.

9 · Drill 3 — Aggregate PlanCh 4 worked · 8 marks

Given: demand=900, 1,100, 1,300; beginning inventory=200; regular production=1,050 each period; ending inventory cost=$4/unit-period.

Decision: the level plan is flow-feasible, finishes with 50 units and incurs $2,800 stated inventory cost before production/workforce costs. Each ending balance becomes the next opening balance.

Exam Numerical feasibility comes before the managerial recommendation. Confirm final position and capacity; then compare relevant categories on the same assumptions.

9b · Plan-Audit GridCh 4 FAQ

9c · Compare Like with LikeCh 4 evidence

Chase=$224,000, level=$248,000 and mixed=$211,000 only under the same demand, limits and final-stock rule. Mixed wins this case, not every case.

10 · Drill 4 — Facility ChoiceCh 5 worked · 8 marks

Given: Site A fixed=$120,000, variable=$7/unit; Site B fixed=$180,000, variable=$4/unit; D=25,000 units. Both have sufficient capacity.

Decision: choose B at 25,000; A is cheaper below 20,000 under the stated assumptions. State the volume range because the recommendation changes at the threshold. Recheck capacity and service.

10b · Fixed-Charge AuditCh 5 model

Two close-cost networks can differ sharply in resilience. A base-case optimum is not automatically implementable.

10c · Recommendation Stress TestCh 5 FAQ

Change demand, cost, capacity loss or service; then compare sites, flows, utilisation and cost. A small base saving may conceal disruption exposure.

11 · Drill 5 — Last MileCh 6 worked · 8 marks

Given: local delivery route=$360/day+$3/stop; parcel carrier=$15/order; one order per stop; sufficient route capacity.

Decision: above 30 daily orders the local route has lower stated cost; below 30 the parcel carrier does. Add capacity, geography and service-quality checks.

11b · Segment the ChannelCh 6 application

Monitor density, participation, failure, response and return cost; name the switch value.

12 · Model Chooser + Trapsall chapters · final recall

High-frequency failure modes

• Using on-hand stock rather than inventory position for ROP.
• Combining raw deviations or mixing weeks with years.
• Calling cycle service and fill rate the same measure.
• Accepting an EOQ outside its price bracket or omitting discount purchase cost.
• Allowing flow when its setup/opening binary is zero.
• Ranking plans with different backlog, horizon or final rules.
• Giving a network winner without a reversal condition.

13 · Two-Sentence Finishreporting rule

Sentence 1: choose the option, give its quantities/flows/cost and comparator. Sentence 2: state the operating reason and reversal condition.