这是一篇来自英国的关于机器学习算法的实验设计与研究的机器智能代写
THE BRIEF/INSTRUCTIONS
The following Learning outcomes will be assessed in this assessment
Use appropriate software tools to develop basic simulations for practical applications.
Develop simple representations and simulations of machine learning processes using appropriate mathematical descriptions and software based simulations.
Assessment Criteria
Weighting (%)
Development of Genetic Algorithm Program 50
Design of Experimental Trials 30
Report 20
Total 100
Introduction and background
In the lectures we have looked at a number of optimisation techniques, one of which was Genetic Algorithms (GA). In this coursework the student will write the necessary software to implement this technique using crossover and mutation operations.
The Scenario
UCLan has designed a new type of diesel engine for use with autonomous vehicles. In order to fine tune and further develop the engine it is important to understand under what circumstances the engine performs best.
Various aspects of the engine can be controlled or modified using additional hardware or tuning, along with this UCLan have developed 5 experimental engine control devices that modify the engines behaviour, these are experimental and we are not fully sure of their effects.
The parameters are as follows :
1) External Temperature (temperature in the engine housing).
2) Internal Temperature (temperature inside the engine).
3) Cylinder Pressure.
4) Value Opening Pressure.
5) Load Torque.
6) NOx Emissions (Nitrous Oxide Emissions)
7) CO Emissions (Carbon Monoxide Emissions)
8) HC Emissions (Hydrocarbon Emissions)
9) PM Emissions (Particle Matter Emissions)
10) UCLan developed Electronic Timing Device.
11) UCLan developed Fuel Flow Device.
12) UCLan developed Emissions Limiter.
13) UCLan developed Battery Management System.
14) UCLan developed Air Flow Management Device.
15) Fuel Injection Timings.
In preparation for using the optimisation methods each of the categories has been converted to a binary input.
Parameters 1,2,3,4,5,6,7,8 & 9 are each represent by a five character binary string. This string indicates whether the parameter is Very Low, Low, Medium, High or Very High.
An instance where the Cylinder pressure was high would be recorded as 00010.
Regarding the five UCLan devices, each device can be fitted to the engine, the engine can be run with 0,1,2,3,4 or 5 devices fitted. The presence of a device is indicted with a 1.
An instance where the engine was run with the Fuel Flow Device and the Battery Management System fitted would be encode as : 01010.
Finally the Fuel Injection Timings have three settings : Low, Medium and High.
An instance of a high injection timing would be encoded as : 001.
This means that that each run can be encoded as a 53 character binary string.
In order to test each run of the engine a fitness function is required, the engine has been modelled and a library (Windows dll) is available on blackboard that will return a grading value for any run given to it. (See Appendix A on how to reference and use the library in Visual Studio for a C# application.)
The library expects to receive a 53 character long String and will return an integer grading rating. The higher the rating, the better the engine has performed.
The student is tasked with :
i) Writing a software program that implement the GA algorithm as discussed in the lecture notes, with the aid of the supplied library.
ii) Designing and carrying out an experimental trial of the UCLan engine to find optimal engine configurations. This should include varied population sizes, mutation rates, varying cross-overs rates and sites etc.
iii)Writing a short report (2000 words max) detailing the program design, the experiment and findings.
Deliverables
- Report, word count not to exceed 2000 words (Not including appendices, tables or figures).
PREPARATION FOR THE ASSESSMENT
The algorithms have been covered in lectures and the lab time during the coursework period will be set
aside for students to develop the coursework program and have access to the lecturer for questions.
程序代写代做C/C++/JAVA/安卓/PYTHON/留学生/PHP/APP开发/MATLAB
本网站支持淘宝 支付宝 微信支付 paypal等等交易。如果不放心可以用淘宝交易！
E-mail: itcsdx@outlook.com 微信:itcsdx
如果您使用手机请先保存二维码，微信识别。如果用电脑，直接掏出手机果断扫描。
