Decision Science: Project Report
This assessment will provide you with the opportunity to:
• Demonstrate your ability to use a simulation model to aid in decision making.
• Develop practical skills in Julia through reverse engineering Julia code.
• Improve your ability to communicate complex facts about simulation results.
• Demonstrate your ability to verify and validate your simulation.
• Practice and improve your skills in data analysis, specifically towards informing a simulation model and analysing the output of a simulation.
This assessment maps to the following course outcomes:
1. communicate how randomness and controlled variation can be used to model complex systems in a range of application domains such as industry, health and transportation
2. create a model of a real-world problem specified in words and implement it as a discrete-event simulation 3. validate results from a discrete-event simulation
4. explore scenarios using simulation to elicit and compare possibilities
5. systematically simulate to derive quantitative information with measures of confidence
6. design simulation-based workflows to support decision-making in real-world contexts.
You are working for Factoriffic Custom Lamps, a manufacturing company. The company produces lamps in an old and inefficient factory—Factory X. You have just been employed to provide them with answers about how to improve the production line at Factory X, and how their current advertising campaign will impact their manufacturing schedules.
You aren’t the first person to have had this remit. Your position is as a replacement for a previous staff member who started the project but left before it was finished. Your task is to continue and adapt their work to help make recommendations for improvements to Factory X, potentially responding to increased demand resulting from a new advertising campaign.
You should treat the exercise as if you are working for your company (rather than doing a student project). The course staff are the “subject matter experts” (or SME) and can answer questions about how the factory works.
The system
The system under study is a small production line with two critical stages referred to here as Machine 1 and Machine 2. Orders for lamps come in and are queued until Machine 1 is free to start manufacture. After being processed in Machine 1, the lamps are queued again, waiting for Machine 2. There is a limited space available for the lamps to be queued (only four orders can wait in this space at most), and if this space is filled, Machine 1 must stop working until there is space. After processing in Machine 2, lamps are shipped to the customer that ordered them.
Hint: The description of the system under study is deliberately vague. Part of your job is to determine a more detailed and accurate model of that system. The course staff will act as subject matter experts so you need to ask them good questions to find out more.
程序代写 CS代考 加微信: cstutorcs
Questions of interest
The company wants to answer the following questions:
1. Is the space available for lamps waiting for Machine 2 sufficient?
2. Could the system be improved if Machine 1 or Machine 2 were made faster?
Note that the company only has enough money to improve one machine and can at most improve speed by a factor of two.
3. What would happen to the current system if the company’s current advertising program were successful, and the rate at which orders arrived increases by 25%?
Would the above improvements help?
Your task is to work through all the steps in a simulation model and prepare a report to make informed recommendations based on the results of your simulation.
Associated files
Download the following files.
• Simulation code: factory_simulation_2.jl
• Example code to run a simulation: factory_simulation_2_run.jl • Measurement data: measured_times.csv
You are provided with Julia code that a previous staff member has written to simulate the system. However, they did not document the code well. That means you have a bit of work to do. The first part will be to conduct VV&A (Verification, Validation and Accreditation) on the simulation.
1. You must build an understanding of the code, and document it. That means you will need to reverse engineer the assumptions, the state model, the events, and entities in the system, as well as any assumptions built into the code. You will document it using techniques described in Module 3: Validation. Document your system state, events and create the various diagrams required.
2. Data has also been provided for the service time distribution for Machine 1 and Machine 2. You will need to use skills developed in Module 4: Including Data into Simulations to analyse this data. You will then need to improve the simulation model and the code to correct any assumptions that need correction.
3. Verify that the code has no bugs and that the output data is sound. Develop a set of unit tests to do so.
In performing these tasks, think of the course staff tutor as subject matter experts who can give you advice about the system under study. Assume they don’t understand simulations, but can help you work on your understanding of the original simulation and how it may or may not be defective.
Hint: Remember to listen to advice from a subject matter expert critically. Ultimately, data is more valuable than advice. Your own determinations should make the best decisions based on all your resources.
In the second part of this assessment, you will use your improved simulation code to explore the system under
study and answer the questions of interest.
1. Write a simulation harness to automate running the simulation code over a range of parameters.
2. Explore the simulation parameters to determine: (i) how long to run each simulation, (ii) how many independent realisations you will use, and (iii) how much burn-in time is needed.
3. Explore the system parameters to determine: (i) what range of parameters you will include in your possibility space, and (ii) with what resolution you will explore those parameters.
4. Use the simulation to generate data that can answer the questions of interest.
5. Analyse the data from the simulations (using what you learned in Module 5) and generate plots to illustrate the answers.
Code Help, Add WeChat: cstutorcs
Having run the simulation, the third part of this project is to write a report from your point of view as an employee of the manufacturing company. The report should be written to your manager, and it should provide convincing results that they can use to make improvements to the production line.
The report should take the be similar to that discussed in Module 6: Reporting and Communicating, and follow the provided template.
It should be concise, but contain enough detail for a reader to understand and be confident about: 1. your verification and validation process, and
2. your results and recommendations.
Ensure that the questions of interest are answered, and that the answers are very clear to a busy reader.
Requirements
You will submit your assessment through Cloudcampus. Save your report as a PDF and upload it to the assessment page. Your work will be assessed on reading your report alone.
The marker will not look at: 1. your code, or
2. your data.
You do not need to include your Julia code or data in the report. Your report must be convincing without access to such details.
It is important to be concise so you will lose marks if the report wastes your manager’s reading time. Assume your manager is very busy and will read the report quickly (and somewhat carelessly).
Consult the assessment rubric for more detail about preparing your submission. Questions can be asked during workshops, consulting or via email.
Grading Criteria
This Project report assessment is worth 25% of your overall grade. This is broken down as
• 5% milestone quiz
• 20% report submission
See rubric below for detailed information on the grading criteria for this report submission. There are a total of 50 marks.
Presentation (6 marks total)
• Overall report structure (1 mark)
– You have used the report template. Material is presented in appropriate sections to help make the
report more readable. • Use of English (2 marks)
– Grammar and spelling are all correct, and stylistically appealing. A formal style is preferred, with some allowed variation for readability.
• Conciseness of report (1 marks)
– Report is clear and expresses what is needed to be said without unnecessary words. Space is used
well and all content (such as figures and tables) is relevant. • Figures and tables (2 marks)
– The use of figures and tables assists in effectively communicating the results of the simulation. All tables and figures are appropriate to the message(s) of the report, are well-captioned, clearly readable and explained in the text.
Executive summary (10 marks total)
• Executive summary (1 mark)
– Summary is clear, precise and concise. It provides concrete answers to the questions of interest.
• Answering the questions (9 marks)
CS Help, Email: tutorcs@163.com
– Provide answers to the three questions posed and also answered the most obvious follow-up questions that your manager might ask. This should include quantitative details from you analysis, expressed in a non-technical way (eg. confidence intervals expressed as percentages).
Documentation (19 marks total)
• Description of system in words (1 mark)
– Report includes a valid description of the system in words.
• A schematic (1 mark)
– Report includes a valid schematic of the system.
• A flow diagram (1 mark)
– Report includes a valid flow diagram of the system.
• A state transition diagram (1 mark)
– Report includes a valid state-transition diagram of the system.
• Modelling assumptions (in a section before the Results) (2 marks) – You have correctly provided the following:
∗ The distributions used for interarrival time and service times for each machine. ∗ Independence between random variables.
∗ Queue input ordering (LIFO, FIFO).
∗ Sizes of available waiting space.
• Verification techniques (4 marks)
– You have provided evidence and discussion of the following:
∗ Unit testing (evidence: results of tests on code functionality)
∗ Integration testing (evidence: results of tests on output files)
∗ Performance testing (evidence: plot showing real-world time versus simulation time)
– Plot: must include a plot of performance testing results • Validation – 1. SME discussion (1 mark)
– Report includes a summary of the discussion with the subject-matter expert. • Validation – 2. Testing assumptions against data (2 marks)
– An appropriate approach is used to test the assumptions against the data and valid results were obtained.
• Validation – 3. Sensitivity analysis of the system with respect to some assumptions (3 marks)
– Report includes a correctly conducted a sensitivity analysis of the system with respect to interarrival
time and service times.
• Exploratory analysis results (3 marks)
– Correctly determine with some discussion a reasonable number of independent simulations and a reasonable length of simulation.
– Correctly determine with some discussion an appropriate burn-in period.
– Plot: must include some relevant plot(s)
Results (15 marks total)
Results: Is the space available for lamps waiting for Machine 2 sufficient? (5 marks)
• (+1) Included plot addresses question of interest
• (+1) Plot shows confidence intervals
• (+1) Discussion of how the displayed results address question of interest
• (+1) Discussion draws reasonable conclusion from the displayed results
• (+1) Discussion includes quantitative information based on simulation output and/or displayed results
Results: Could the system be improved if Machine 1 or Machine 2 were made faster? (5 marks)
• (+1) Included plot addresses question of interest
• (+1) Plot shows confidence intervals
• (+1) Discussion of how the displayed results address question of interest
• (+1) Discussion draws reasonable conclusion from the displayed results
• (+1) Discussion includes quantitative information based on simulation output and/or displayed results
Results: What would happen to the current system if the company’s current advertising program were successful, and the rate at which orders arrived increases by 25%?(5 marks)
• (+1) Included plot addresses question of interest
• (+1) Plot shows confidence intervals
• (+1) Discussion of how the displayed results address question of interest
• (+1) Discussion draws reasonable conclusion from the displayed results
• (+1) Discussion includes quantitative information based on simulation output and/or displayed results