BMGT43660 Decision Making Strategies and Techniques Individual Assignment 2026 | UCD

BMGT43660 Individual Decision Analysis Assignment

(15% of module grade)

Task

• Answer parts (a) and (b) of the problem presented in this document.

Instructions

1. You will get a lower grade if you don’t read and follow all the instructions on this page and the next – so keep reading!
2. Your decision tree and risk profiles can be either hand drawn (in which case they must be very neat and clear) or drawn with the aid of a computer (such as the examples in lecture slides).
3. The report should include only one decision tree. That tree should demonstrate a correct structure for the problem.
4. Consequence values should be shown at the right-hand side of the end branches of the decision tree.
5. Calculations relating to consequences, probabilities, expected values, and risk profiles must be included in your report. Credit cannot be given for elements of your report not backed up by the relevant calculations.
6. Any diagrams, charts, or tables must be clearly and fully visible within the document. Do not try to submit charts and tables as separate files and do not submit the document or any part of it as a web-linked file.
7. If you are including any hand-drawn material, you will need to ensure that it is cleanly scanned and fully legible. If you submit material that is poorly legible or not entirely visible you will not be given another chance for submission. Your work should be neat, clear and well-presented (see Assessment Criteria at the end of this document).
8. There is no page count or word count requirement for this assignment.

Deadline and Submission

1. The deadline for submission is 08:59am on Wednesday 8 October.
2. You must submit your assignment through Brightspace. Hard copy or soft copy submitted in class cannot be accepted. You can submit only one document. You can submit your document more than once on Brightspace up to the deadline, and only your most recent submission prior to the deadline will be received.
3. The assignment must be submitted as a single PDF document. It will not be possible to upload multiple files or any other file type. If you have worked on the assignment in multiple files/documents, you must compile them into a single PDF document prior to submitting your work.
4. Submissions received after 08:59am on Wednesday 8 October will be penalised at a rate of one grade (up to one week late) or two grades (up to two weeks late). It will not be possible to submit more than two weeks late, and students who have not submitted by then will not receive a grade or credit for this assignment. Upload your completed assignment well in advance of the deadline, to avoid any last-minute problems.
5. The report must be your own original work. In your submitted assignment, you must copy and paste the Assignment Submission Form provided, including your signature at the end of the form. Collaboration, assistance, cross-checking, comparison or sharing with other students is prohibited. Do not discuss or share any aspect of this assignment with another student. In the event of any form of collaboration, sharing, or plagiarism, the module coordinator will penalise ALL students involved by giving a zero grade or reduced grade, and/or may opt to refer those students before a School-level plagiarism committee with serious potential consequences.

QZX Decision Problem

QZX makes hi-tech scientific instruments and is developing an AI-assisted microscope. To succeed, they need a specialised AI chip, which QZX can try to develop by either working alone or in partnership with another organisation.

If QZX decide to go it alone, the NPV^* of the project will be €60 million. There are two possible partners, and QZX has collaborated with both in the past: Oxbridge Micro or Weiss Medica. Working with Oxbridge would improve the chance of success but reduce the NPV for QZX to €45 million. Working with Weiss entails an even greater chance of success but reduces the NPV for QZX to €30 million.

QZX believes that if they go it alone, they have a 25% chance of producing the AI chip. With Oxbridge, the chance of success would be 42%, and with Weiss, 80%. They can’t work with both, and must decide right now whether to go it alone or work with a partner. They can’t switch or adopt a partner later.

The prototype

Whether alone or with a partner, in 12 months QZX will produce a prototype AI chip. The prototype will be tested for viability. If it is viable, it is more likely to lead to successful completion of the AI chip. However, even a viable prototype does not mean successful completion of the AI chip is guaranteed, and a nonviable prototype does not mean that failure is inevitable.

A prototype produced by QZX alone would have only a 20% chance of being viable. Data from past partnerships with Oxbridge and Weiss can be used to find the chance of partnership leading to a viable prototype. All partnerships between QZX and Oxbridge or Weiss produced prototypes, and half of those prototypes passed the viability test. Oxbridge was the partner in half of those viable results,while Weiss was partner for the other half of those viable prototypes. The remaining half of prototypes produced in partnership were non-viable; Oxbridge was partner for 75% of these non-viable prototypes, and Weiss for only 25%.

Completing the AI chip

After the prototype stage, efforts to complete the AI chip will take another year. The probability of success in completing the AI chip is better if the prototype is viable than if it is not viable, as shown below:

If they fail to complete the AI chip, they can’t try again: QZX would have to release a microscope without the AI chip, which would lead to a reduced NPV. In the event of such failure, the NPV would be €10 million if working alone, –€5 million (a loss) with Oxbridge, and –€15 million (a loss) with Weiss.

A future alternative

NVIDIA have announced that in one year, they will release an AI chip for use by other companies. The NVIDIA chip will definitely be available in one year and would guarantee the success of the project. However, using it would significantly reduce the NPV.

If QZX stop their development efforts in one year and buy the NVIDIA chip, the NPV if QZX is working alone would be €18 million, while with Oxbridge it would be €14 million, and with Weiss, €10 million. Alternatively, they could just wait one year without making any effort to develop an AI chip, and then use NVIDIA; this would yield an NPV of €20 million.

If they fail to develop their own AI chip after two years, it will be too late for QZX to buy the NVIDIA chip, because the market and technology will have moved on. As indicated above, they would have to release a product without an AI chip.

Questions

(a) Draw a decision tree that represents the whole decision problem. You may present only one decision tree. Solve using the rollback process. Assume that the decision is based solely on NPV. Make a recommendation based on your findings. In your accompanying report, show all of your consequence, probability, and expected value calculations. (65 marks)

(b) Prepare and compare cumulative risk profiles for four scenarios:

• the best decisions QZX could make if working alone
• the best decisions QZX could make if working with Oxbridge
• the best decisions QZX could make if working with Weiss
• waiting one year, and then using NVIDIA

Draw conclusions and make recommendations, commenting on whether any of the risk profiles are dominated by any of the others. In your report, show all preparation and calculations for the risk profiles (35 marks)