ACS6124 Multisensor and Decision Systems
Multisensor and Decision Systems
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: THEend8_
ACS6124 Multisensor and Decision Systems
Part II: Decision Systems for Engineering Design
Assignment
ACS6124 incorporates two assignments – one for each Part of the module. This document
introduces the assignment for Part II, providing submission instructions, a detailed assignment
briefing, and the marking criteria.
Assignment weighting: 50% of the module
Assignment released: 25 April 2022 (Monday, Week 9)
Assignment due: 12 noon, 23 May 2022 (Monday, Exam Week 1)
Format: A report of 15 pages maximum (using a top and bottom margin of 1.5
inches, a left and right margin of 1 inch, text of size 12 point, with 1.5
line spacing). The report must be submitted electronically via
BlackBoard.
Assignment code: ACS6124-002
Penalties for late submission
Late submissions will incur the usual penalties of a 5% reduction in the mark for every working day
(or part thereof) that the assignment is late and a mark of zero for submission more than 5
working days late. For more information
Unfair means
This is an individual assignment. You should not discuss the assignment with other students or
work together with other students in its completion. The assignment must be wholly your own
work. References must be provided to any other work that is used as part of the assignment. Any
suspicion of the use of unfair means will be investigated and may lead to penalties.
Extenuating circumstances
If you have any medical or special circumstances that you believe may affect your performance on
the assignment then you should raise these with the Module Leader at the earliest opportunity. You
will also need to submit an extenuating circumstances form.
Feedback
Written feedback will be provided on Blackboard within 15 working days, in line with Department
guidelines.
2
Assignment briefing
Imagine you are a recent graduate who has decided to build a start-up company working on
decision systems for engineering design. Your first potential client is a company working on a new
type of nuclear-powered vehicle. While their team is capable of designing the new powertrain, they
have limited experience in designing the rest of the vehicle; for example they need to implement a
suitable controller for their propulsion system and are not sure how it might affect the dynamic
performance of the vehicle.
Your task is to convince them that your start-up company is perfectly suited to help them in the
decision making process. To do that, they have asked you to prepare a report to highlight the
capabilities of the tools you are proposing to use. Because at this stage they are reluctant to share
their Simulink models, your pilot study will focus on tuning the gains for a Proportional-Integral (PI)
controller, such that a feedback control system satisfies a set of requirements.
The system to be controlled, and the performance criteria against which a set of controller gains
are assessed, are described in the Laboratory A instructions. The goals for the performance
criteria are given in the instructions for Laboratory B.
During Laboratory A, you will learn about the relationships between the design variables and the
performance criteria for the given system. In Laboratory B, you will attempt to optimize the gains to
meet the goals for the performance criteria.
Following these laboratories, you need to write a report that would appeal to both the CEO of your
potential client and their Chief Engineer. As you write, you will have further time to explore the
system and perfect your design in the open Laboratory C. You need to convince the CEO that
multi-objective optimization is the best way to approach the decision making. So far her company
have used opinions from experts and developed prototypes to validate their designs. In addition,
you need to have a technical part in which you show their Chief Engineer what your pilot study has
managed to achieve, explaining any challenges encountered in satisfying all the requirements, and
making recommendations for tuning options.
Your report should be structured as follows:
Title page including Executive Summary (1 page)
Summarise the outcomes of the tuning process and recommendations for PI gain settings
in under 300 words. This section does not contribute to the page limit for the report.
Section 1: Multi-objective optimization for Engineering Design (3 pages)
Write a brief introduction to Decision Systems for Engineering Design. Explain how it
compares with other approaches used in decision making and give five examples from the
literature where it has been used for vehicle design. Draw a comparison between three
classes of population-based optimizers that can be used as the engine for a multi-objective
optimization process. Explain the main differences in their approach to find a candidate
approximation set.
Section 2: Problem Formulation (1 page)
Express the problem in formal mathematical terms.
Section 3: Sampling Plan (2 pages)
Show at least three different sampling plans and analyse their space-filling performance.
Identify a sampling plan to take forward.
3
Section 4: Knowledge Discovery (2 pages)
Use the evaluations from the chosen sampling plan to describe the relationships between
the design variables and performance criteria.
Section 5: Optimization Process (2 pages)
Describe the optimization approach used and how goals were incorporated into the
process.
Section 6: Optimization Results (2 pages)
Show the results of the optimization process, indicating whether or not the goals have been
met, and the trade-offs inherent to the problem.
Section 7: Recommendations (1 page)
Based on the knowledge discovery and optimization results, make recommendations for PI
controller options for consideration by the Chief Engineer.
Section 8: Conclusions (1 page)
Link the results of your study with the vehicle propulsion problem your client is keen to
solve. How would you apply the same methodology for their problem? Indicate at least two
other decision systems tools that you propose to use to help them in their design problem.
Bibliography
Include references to any works used in the report. This section does not contribute to the
page limit for the report.
Appendix
Provide your Matlab code listings as an appendix to the report. The appendix does not
count towards the page count for the report.
4
Marking criteria
The assignment will be marked out of 100. The marking criteria below provide guidance on the
relationship between the quality of submission and the marks awarded. Note that the quality
statements are indicative only – the actual mark awarded will be a holistic judgment of the overall
quality of submitted work.
Mark
awarded
Expected attributes of the technical report
70-100 • An executive summary that succinctly summarises the findings of the tuning process and
recommendations for future action.
• A coherent introduction to decision systems for engineering design, contrasting multi-
objective optimization to other approaches to decision support. Accurate description and
comparison of the three major classes of multi-objective optimizer.
• A problem formulation that correctly interprets the problem features in the language of
constrained multi-objective optimization, including identification of design variables,
parameters, objectives and constraints.
• A set of at least three sampling plans that have been correctly assessed in terms of their
space-filling properties.
• Appropriate and creative data mining and visualisation of the sampling plan evaluation,
identifying key relationships between design variables and objectives (e.g. regions of
stability, trade-offs between aspects of transient performance).
• A clear description of the optimization approach used, including how Chief Engineer
preferences were incorporated into the search process.
• Appropriate and creative data mining and visualisation of the results of the optimization
process, identifying the level of success achieved and areas of conflict that are as yet
unresolved.
• A coherent and credible set of recommendations for the controller gain settings, reflecting
the results of the knowledge discovery and optimization processes.
• Compelling association of the study findings to the client’s vehicle design problem, indicating
how the same methods could be used to deliver benefits, and highlighting two other
decision systems tools that would be used alongside these methods.
• Well-presented report, with appropriate use of labelled figures and few spelling or
grammatical errors.
60-69 • An executive summary that succinctly summarises the findings of the tuning process and
recommendations for future action.
• An introduction to decision systems for engineering design, and accurate description of the
three major classes of multi-objective optimizer.
• A problem formulation that correctly interprets the problem features in the language of
constrained multi-objective optimization, including identification of design variables,
parameters, objectives and constraints.
• A set of at least three sampling plans that have been correctly assessed in terms of their
space-filling properties.
• A creditable attempt to identify key relationships between design variables and objectives
through visualisation of the sampling plan evaluation (e.g. regions of stability, trade-offs
between aspects of transient performance).
• A clear description of the optimization approach used.
• A creditable attempt to analyse the results of the optimization process, identifying the level
of success achieved.
• Recommendations for the controller gain settings that are largely grounded in the results of
the knowledge discovery and optimization processes.
• Linkage of the study findings to the client’s vehicle design problem, highlighting at least one
other decision system tool that would be used alongside these methods.
• Generally well-presented report, with appropriate use of labelled figures and few spelling or
grammatical errors.
5
Mark
awarded
Expected attributes of the technical report
50-59 • An executive summary that includes an attempt to summarise the findings of the tuning
process and makes recommendations for future action.
• An introduction to decision systems for engineering design, with an accurate description of
at least one of the classes of multi-objective optimizer.
• A problem formulation that interprets the problem features in the language of constrained
multi-objective optimization, but where the formulation may contain some missing or
unclear elements.
• An appropriately visualised sampling plan.
• Some attempt to identify key relationships between design variables and objectives through
visualisation of the sampling plan evaluation (e.g. regions of stability, trade-offs between
aspects of transient performance).
• A description of the optimization approach used, although some aspects may not be clearly
described.
• A creditable attempt to analyse the results of the optimization process, identifying the level
of success achieved.
• Recommendations for the controller gain settings that are largely grounded in the results of
the knowledge discovery and optimization processes.
• Some indication of how the study findings link to the client’s vehicle design problem,
indicating how the same methods could be used to deliver benefits, highlighting at least one
other decision system tool that would be used alongside these methods.
• Generally well-presented report, with appropriate use of labelled figures and few spelling or
grammatical errors.
40-49 • An executive summary that provides a readable summary of the report, but is lacking focus
on findings and recommendations.
• An introduction to decision systems for engineering design, with a description of at least one
of the classes of multi-objective optimizer.
• A problem formulation that interprets the problem features in the language of constrained
multi-objective optimization, but where the formulation may contain some missing, unclear
elements, or incorrect elements.
• An appropriately visualised sampling plan.
• Lacking a convincing analysis of the key relationships between design variables and
objectives through visualisation of the sampling plan evaluation (e.g. regions of stability,
trade-offs between aspects of transient performance).
• A description of the optimization approach used, although some aspects may not be clearly
described.
• Results of the optimization process are presented, but these are not analysed.
• Lacking recommendations for the controller gain settings, or recommendations that do not
relate to the results of the knowledge discovery and optimization processes.
• Lacking indication of how the study findings link to the client’s vehicle design problem,
although highlighting at least one other decision system tool that would be used alongside
the methods employed.
• Issues with the presentation of the report, with numerous grammatical errors and figures
that are missing labels.
6
Mark
awarded
Expected attributes of the technical report
0-39 • Missing or incoherent executive summary.
• Lacking an introduction to decision systems for engineering design and/or substantial
inaccuracies in the description of multi-objective optimizers.
• Missing or incoherent problem formulation.
• Some evidence of a sampling plan, but unclear what this looks like.
• Lacking a convincing analysis of the key relationships between design variables and
objectives through visualisation of the sampling plan evaluation (e.g. regions of stability,
trade-offs between aspects of transient performance).
• Missing or incoherent description of the optimization approach used.
• Missing the results of the optimization process.
• Lacking recommendations for the controller gain settings, or recommendations that do not
relate to the results of the knowledge discovery and optimization processes.
• Lacking indication of how the study findings link to the client’s vehicle design problem and
absence of consideration of other decision system tools that could be used alongside the
methods employed.
• Major issues with the presentation of the report, with numerous grammatical errors and
figures that are missing labels, such that the meaning in the report is hard to discern.