SCM 7320 Case analysis guidelines
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
Larsen & Toubro – Spare Parts Forecasting
Case analysis guidelines – SCM 7320
Overview
The purpose of all the case studies in this course is to give you practice doing the following:
• Framing a problem
• Building a model
• Analyzing the results of the model and testing alternatives
• Communicating the results ofyour analysis and making a recommendation
Approach
Start with framing the problem: don’t immediately jump ahead to “we need to make a forecast”. Instead consider what decision the manager is trying to make. The case states that company is having difficulty balancing the cost of maintaining an appropriate level of inventory while avoiding unavailability of spare parts due to shortages. In addition, there are over 20,000 items to track, making modeling even more challenging.
1. What information do you need to make for those decisions? Is all that information available in the case study? Document this in your summary.
Modeling
2. What strategy should Vijaya Kumar adopt for developing forecasting models for demand estimation ofthe 20,000 spare parts in inventory?
The case provides month-wise demand data for 11 items. Some ofthese items have a high demand, while others can go several months without any demand. Further, the value ofthe items (and hence the cost of holding inventory) varies widely. Finally, the items vary according to the A-B-C categorization of % sales, with A representing the largest fraction as opposed to the largest value.
For the purpose of this analysis you should focus your efforts on the following parts:
• Item No. 07063-51210I1
• Item No. 6735-51-5143I
• Item No. 600- 181-6740I
• Item No. 198-27-42263
3. Develop several forecasting models for each ofthe parts listed above. To make the forecasts you will want to consider several different approaches:
- Linear regression
- Weighted moving averages
- ARIMA methods (if discussed previously)
- Various exponential smoothing methods
Analyzing
You will need to decide which forecast model is best, based on some objective measures of forecast accuracy, such as: MAD, MSE, MAPE, AIC, etc.
4. Do the forecast errors appear to be uniformly distributed? In other words, when you plot the residuals do you see any trends? What should you do ifthere are trends in the errors?
5. Which forecasting method would you recommend for each ofthe spare parts you analyzed? Are there any underlying assumptions that could change the results?
Communicating
Your case analysis should include the following:
Summary
- Frame the problem you are analyzing, including the answers to the first question
- Briefly describe the model(s) you used for your analysis
- State your conclusions/ recommendations
Analysis
- Describe which models you developed for the annual and monthly forecasts. You should also include any plots or graphs of the data that you made.
- Summarize the results of at least 2 different models for each forecast in a table that includes all important parameters and measures of accuracy.
- Provide your answers to questions 4 and 5
Recommendation
- Clearly state your recommendation(s) for what the manager should do. How will the manager know ifyour recommendations are successful? What metrics would you track to determine this?
- Provide an economic evaluation ofyour recommendation, if applicable.
Appendix
- Include the R-script files you used to analyze the problem and any other data as either tables or exhibits that you referenced in your analysis.
2022-02-14