OMGT2229 Strategic Supply Chain Management Assignment 3 Sample (RMIT) Singapore

For information and reference only – not to be submitted as your own work.

This extract shows how the quantitative parts of RMIT University OMGT2229 Strategic Supply Chain Management, Assignment 3 (Individual Quantitative and Qualitative Case Study, 40 marks) can be worked and presented. The case follows a Harvey Norman Singapore scenario. The original brief is not yet on our solved questions hub, and we will add a direct link when it goes live. For marking advice, read the companion post on how to approach OMGT2229 Assignment 3.

The original assessment brief is available on our question page: OMGT2229 Strategic Supply Chain Assignment 3: Harvey Norman Forecasting, EOQ and Sourcing Decision Tree.

The brief has two tasks. Task 1 (20 marks) forecasts monthly demand for a Wi-Fi security camera, compares two methods and applies EOQ with quantity discounts. Task 2 (20 marks) compares sourcing options with decision trees. The infographic and reflective journal are your own work, so only guidance is given for them below.

Task 1: demand data and assumptions

Monthly demand for the camera runs from January 2025 (425 units) to June 2026 (650 units), 18 observations in total. All figures below were computed in Python and rounded for display. Keep the full-precision values in your Excel tabs so that your report and workbook agree.

Demand shows a clear upward trend. With only 18 months, a 12-month seasonal cycle cannot be estimated reliably, because each month would have just one or two observations. The extract therefore assumes a six-month cycle (three complete cycles). This is one defensible approach, not the only one. State your chosen periodicity and your reason in the report.

(a) Three-month moving average forecast, July to September 2026

The forecast is the mean of the latest three actual values: (725 + 615 + 650) / 3 = 663.3 units. A simple moving average gives the same value for each future month, so July, August and September are each forecast at about 663 units.

(b) Static method forecast, July to September 2026

The static method deseasonalises demand, fits a linear trend, then re-applies seasonal indices. The steps are as follows.

1. Compute a centred six-month moving average for months 4 to 15 (the half-weight rule applies at the window ends because the period length is even).
2. Divide actual demand by this average to get a seasonal factor for each month. Average the factors that share a position in the cycle, then rescale so the six indices sum to 6.
3. Divide demand by its index to deseasonalise it, then regress deseasonalised demand on time (t = 1 to 18).
4. Forecast = (intercept + slope x t) x seasonal index.

The regression gives intercept = 462.65 and slope = 12.12 units per month. July 2026 is t = 19 (position 1), August is t = 20 (position 2) and September is t = 21 (position 3).

(c) Forecast accuracy: MAD, MAPE and tracking signal

To compare the methods fairly, both are tested on the same 15 months (months 4 to 18), the first month for which a three-month average exists. The error is actual minus forecast. The tracking signal is the running sum of errors divided by the running MAD. The static fits are in-sample, because the seasonal indices and trend use all 18 points. Say so in your report as a limitation.

(d) Preferred method

The static method is preferred. It has the lower MAD and MAPE, and its tracking signal stays within a common +/-4 control limit. The moving average lags a rising series, so its errors are mostly positive (demand above forecast) and its signal climbs to 7.14. That pattern signals bias, not random error. The static method’s weakness is its reliance on a six-month cycle estimated from three repeats, so review the indices as new data arrive.

(e) EOQ and quantity discounts

Annual demand is D = 640 x 12 = 7,680 units. Ordering cost is S = $2,500 per order and holding cost is 15% of unit cost, so H = 0.15 x price.

EOQ without discount (price $70). H = $10.50. EOQ = sqrt(2 x 7,680 x 2,500 / 10.50) = 1,912 units, or about 4.0 orders a year.

Total annual cost at $70. Purchase cost = 7,680 x 70 = $537,600. Ordering cost = (7,680 / 1,912.4) x 2,500 = $10,040. Holding cost = (1,912.4 / 2) x 10.50 = $10,040. Total = $557,680.

Next, test each price band. If the EOQ falls outside its band, the best order size in that band is the nearest band limit.

The lowest cost is at $69 with Q = 2,501 units. The annual saving against the no-discount policy is 557,679.84 – 550,539.60 = $7,140.24. Notice that the $68 band is cheaper per unit but loses on holding cost, which is the usual lesson in all-unit discount problems.

Task 2: sourcing decision trees

Assumptions stated up front

• Illustrative base demand is 7,500 units a year. Each period, demand is 8,250 (+10%) or 6,750 (-10%), each with probability 0.5.
• The China price is 385 CNY at a base rate of 5.5 CNY per SGD, which is 70 SGD. “CNY strengthens 5%” is read as fewer CNY per SGD (5.225), so the SGD cost rises to 73.68. “CNY weakens 5%” is read as 5.775 CNY per SGD, so the cost falls to 66.67.
• Exchange-rate moves are applied independently to the base rate in each period. If your unit guide compounds them, state that and recompute.
• Selling price is 120 SGD. The discount rate is k = 10% per period, and period 0 is excluded, so NPV = E1 / 1.1 + E2 / 1.1^2.

Tree logic

For each option, draw a decision node, then a demand chance node, then an exchange-rate chance node. That gives four branches per period (probability 0.25 each) and 16 paths across two periods. Put the period profit on each branch, then roll back using expected values.

Worked example, China, demand 6,750 and CNY strong: revenue = 6,750 x 120 = 810,000; cost = 8,500 x 73.68 = 626,316; profit = 183,684. The unsold 1,750 units are donated, so they earn nothing.

Expected profit and NPV

The discount factor sum is 1/1.1 + 1/1.21 = 1.7355. China and local are almost equal on expected value, but China carries far more spread (standard deviation of period profit about 94,800 against 30,000 for local). The hybrid wins on both return and risk under these assumptions, because it commits to fewer units up front and pays the 90 SGD premium only on the shortfall. Test this with sensitivity runs, such as a higher local top-up price, before you recommend it.

Guidance for the infographic and reflective journal

The infographic and journal must be your own work. For the infographic, pick three or four forecasting practices used by consumer electronics retailers (for example, sales history, promotions calendars, supplier lead-time signals) and show how each feeds a decision. For the journal, describe what the numbers changed in your thinking and cite at least one source. Use the Harvard style your course requires.

Need help checking your own workbook against these methods? Our operations management assignment help team can review your logic, and our report writing help covers structure and referencing.

About the expert

Renshu Phua holds a master’s degree in Business Analytics from the National University of Singapore. The profile lists 8 years with Singapore Assignment Help and 2,322 completed orders, and describes numerous essays, case studies and reports on business management.

References

• Chopra, S. and Meindl, P. (2019) Supply chain management: strategy, planning, and operation. 7th edn, Global edn. Harlow: Pearson.
• Harris, F.W. (1990) ‘How many parts to make at once’, Operations Research, 38(6), pp. 947-950. (Reprinted from Factory, The Magazine of Management, 1913.)
• Hyndman, R.J. and Athanasopoulos, G. (2021) Forecasting: principles and practice. 3rd edn. Melbourne: OTexts. Available at: https://otexts.com/fpp3/ (Accessed: 19 October 2026).
• Hyndman, R.J. and Koehler, A.B. (2006) ‘Another look at measures of forecast accuracy’, International Journal of Forecasting, 22(4), pp. 679-688.