Assignment: Case Background
A manufacturing company has the following areas of concern and has hired you as a Consultant to analyse them and write a technical report in order to help the manager make informed decisions.
Area 1: Decision trees
This company is considering expansion of its current capacity, the options are to build a small, a medium or a large plant. Of course, there is also the option of doing nothing. The new plant would introduce a new product and potential marketability of that product is relatively unknown, forecasts say it is 70% likely that the market will be unfavourable. If a large plant is built and a favourable market exists, a profit of £110,000 would be realised. However, an unfavourable market for this plant would yield £70,000 loss. If a medium plant was built it would generate £50,000 profit within a favourable market, but a loss of £8,000 would result from unfavourable market. A small plant, on the other hand, would return £30,000 with favourable market conditions and lose only £3,000 in an unfavourable market.
There is an option to get hold on an updated forecast that could provide some key information for the decision. The manager knows, from previous experience, that the reliability of the updated forecast in relation to the original forecast, is as follows.
Area 2: Data mining
Last month the marketing department sent a mailing to 40,000 of the company’s existing customers with a special offer. The response was exciting: 2% of them responded, which brought in £60,000 in revenue. The manager is now considering continuing the program, with a budget of £20,000, which will allow the marketing department to target another 20,000 customers (out of your customer base of 100,000). The manager does not want to just target them randomly, as previously done. So, the idea is to classify the customers into two categories of responsive and non-responsive to this offer, and only target those from the responsive category. The manager has asked you to build two data mining models for the classification purpose, i.e., a tree model and a K-nearest Neighbours model. Therefore, you need to evaluate them before deciding which model is preferable.
Area 3: Linear Programming
There are four different types of products made at the company: Essentials, Comfort, Practical and DeLuxe. In order to meet demand, the company needs to make at least 900 Kg. of Essentials, between 200 and 300 Kg. of Comfort, no more than 250 Kg of Practical and no more than 100 Kg. of DeLuxe. Each product Kg. of the products, contains 40%, 70%, 50%, 90% of Component 1 with the remaining weight made of Component 2. The company has 1200 Kg. of component 1 and 700 Kg. of component 2 for use next week. Each machine has 48 hours of time available next week.
The following table shows the time required for each product on each machine:
Minutes required per Kg.
The financial manager has presented the following:
Selling and Administrative
Allocated fixed costs
Selling and Administrative
Net Profit per Unit
Prior to conducting any analyses, you meet with the manager at the company to verify what are the aims and objectives of the project-report that you need to produce.
Meeting with the manager
The manager tells you that the report should be concerned with three equally important areas:
Area 1: Decision trees. The manager wants you to provide a structured approach and a graphical representation to explain the expansion problem, first with the original forecast to calculate the expected monetary values for each option, and to provide recommendations based on your analysis. Second, the manager wants you to consider the updated forecast information and to help them decide what is the value of the imperfect information. Finally, to recommend a course of action.
Area 2: Data mining. The manager wants you to present and describe the confusion matrix and explain how you will fill it out for one of models. Then, present and describe the cost/benefit matrix for this case. Also, justify why these matrices are important for evaluating data mining models. Next, show the evaluation function you will use to compare the models. Finally, discuss how the confusion and cost/benefit matrices come into play in this evaluation function.
Area 3: Linear Programming. The manager wants you to conduct a thorough Linear Programming analysis including sensitivity analysis using a spreadsheet modelling software such as MS Excel Solver. The manager wants to know in particular: (a) What is the optimal production of each type of product they should make in order to increase net profit per unit and (b) Explain the meaning of the shadow price results for two results in the sensitivity report.
The manager wants you to write a report to support your analyses and findings. Your report to the manager should be structured to include the following points:
Brief introduction. This should detail the purpose and structure of the report.
Research Strategy. This section should include but not be limited to, the following for each area of the project:
plan, outline and justification of the modelling method used
implementation issues (e.g. modelling assumptions)
Data analysis. This section should include but not be limited to, the following discussions for each area:
relevant graphical and/or tabular output
description of concerns raised with the data and the models used
description of how you determined the types of model(s) to use
Discussion of the results. This section should include but not be limited to, the following for each area:
summary presentations of the findings from your analyses
the benefits/downsides of using your selected models for this case study
any insights or suggestions based on your findings
References: provide some evidence of reading, all referencing must be in Harvard format (look beyond the module notes)
Appendices (only required if necessary).
Summarize your work, using selected MS Excel output, in a word-processed report with all appropriate results and analyses. Note: you should only include important output such as charts, graphs and tabular data that support or enhance your report. You should not include all output generated; you must be selective. Summarising output in your own tabular format is acceptable.
Your report should be no more than 2500 words; this limit excludes tables, figures, captions, table of contents, references and appendices. Beyond this, if you are not sure then assume it is included in the word limit. Your report should be written using size 12 Arial font, justified and use double line spacing (this lends itself to easier reading).
Divide your report into sections with clear headings (use the format described in points 1-6), summarise key points up front. You should provide a clear report title. Pay attention to the professional appearance and use of language. Do not use ‘clip art’ or fancy fonts and page borders. Figures and tables should be legible, correctly titled and include a caption. Avoid the use of footnotes and tabloid style presentation. Show correct use of grammar with no spelling errors. Use appendices only if necessary and choose to use them for their relevance and not as a continuation of the debate (or they will be included as part of the word count). The style of the report should be appropriate to a management report and NOT like an essay.
END OF ASSESSMENT TASK