辅导COMP6216、辅导Python/Java程序
- 首页 >> Algorithm 算法 Coursework outline for COMP6216 – Simulation Modelling
for Computer Science
The assessment consists of two components:
Coursework Assignment I (worth 30%)
Give a 10 minute talk (+2 minutes questions) about a simulation modelling paper published in a peer
reviewed journal (see slides from first lecture – http://users.ecs.soton.ac.uk/mb8/sim2020/Intro.pdf --
for a list of suggested papers). The talk should give
(a) a brief overview of the area of research the paper addresses,
(b) explain its contribution to the area, and
(c) give a brief overview over the type of simulation modelling being used.
Marks will be given on:
(i) if the delivery of the talk is according to a standard that it could be used in teaching/presented at a
conference, (ii) your comprehension of the paper and how well you answer questions, (iii) if your slides
meet professional standards, and (iv) the general organisation of the talk and how well you covered the
aspects mentioned above. In the assessment, the criteria (i)-(iv) will have equal weight.
Talks will be scheduled after Easter, so have your talk ready at the end of Easter break, it can be
scheduled in any lecture/seminar slot after that date.
Coursework Assignment II (worth 70%)
A modelling problem is described in the first set of lecture slides (see material copied from the lecture
slides below). You are to address this problem using two different modelling techniques: first, using a
differential equation-based approach; second, using agent based modelling. Compare your findings
from both approaches and write a 6 page conference paper that summarises your findings. The 70%
marks you can obtain for this part of the coursework are split as follows:
Quality of the writing and figures in the report 10%
Technical work:
Develop a model based on differential equations addressing the problem;
that is, give a differential equation that models the problem, describe how
you derive it, and reason about its classification (7.5%).
40%
Explore the use of analytical techniques to gain insight into the system's
behaviour; that is, analyse the above differential equation, find full (or
equilibrium) solutions, and argue about their stability (7.5%).
Numerically integrate the differential equations using an appropriate
integration method of your choice and compare the results to the
analytical results; that is, develop the computer code for an integration
scheme (from scratch, do not use off the shelf libraries or integrated
functions provided by math software) and give evidence that it
reproduces system behaviour correctly, and reason about parameter
choices made in your integration scheme (10%).
Implement an agent-based model that addresses the same problem. Give
evidence (for example, appropriate example output) of the model you
have implemented, and compare your findings to the results obtained
with other methods (from above) (15%).
Use both models to answer the research questions given below.
Quality of an original extension of the problem:
Marked according to originality (Is this an interesting question to ask in
this context), quality of the analysis (do you understand what is going on
and did you use appropriate techniques to analyse the model), and its
motivation (can you convince me that it makes sense).
20%
Total for CW assignment II 70%
Reports are due at the end of term (May 15) and will have to be submitted in electronic form as a pdf
file. Don't forget to upload any simulation code you developed for the assessment.
Further instructions about the assessment can be found in the first set of lecture slides (copied into the
document here):
Problem description from lecture slides
Consider the following situation:
● A population of students is working on group projects. Students can follow two strategies (S): work
hard for the project (S=H) or free-ride (S=L).
● In every course, groups of size n are formed at random. Students use the strategy determined at the
beginning of the course (i.e. S=H or S=L) in their group work.
● Total group effort is determined by the composition of the group. In a group with h hard workers and
l=n-h lazy workers total group effort is e=h*H+l*L (H and L being the effort put in by hard/lazy
workers).
● When group projects are marked, every student in a group gets the same mark. The lecturer
determines this mark as m=e/n (that is, by dividing total group effort by the number of students in a
group; the larger this number the better the mark.)
● At the end of the semester groups are dissolved and every student rethinks his strategies for the next
semester. He/she does this by selecting another student (a reference student) at random and comparing
a measure ? based on marks and effort, =m-a*S (where a is a parameter and S=H or S=L depending
on strategy). The measure accounts for the mark obtained, but is lowered by the amount of effort spent.
Very good marks without effort maximise the performance measure. If the reference student selected
for comparison got a higher performance measure, the selected student will imitate the reference
student's strategy in the next semester with a probability that is proportional to the difference in the
performance measure. Students will not imitate strategies from other students with worse
performance measures.
● Students study forever (that is, take an infinite number of courses) and follow the same procedure (as
outlined above) for every course they take.
Research questions to address:
● Assuming we start with equal numbers of hard working and lazy workers, what is the composition of
the group
– After 4 years (i.e. 8 courses) if H=1 and L=0 and a=0.5.
– In the long run (after an “infinite” number of years)?
– How quickly is this equilibrium state reached?
● How do the following parameters influence results:
– Initial composition of the population
– Group size (n)
– Cost of effort, a
– Contribution of hard workers to group effort (i.e. H and L)
for Computer Science
The assessment consists of two components:
Coursework Assignment I (worth 30%)
Give a 10 minute talk (+2 minutes questions) about a simulation modelling paper published in a peer
reviewed journal (see slides from first lecture – http://users.ecs.soton.ac.uk/mb8/sim2020/Intro.pdf --
for a list of suggested papers). The talk should give
(a) a brief overview of the area of research the paper addresses,
(b) explain its contribution to the area, and
(c) give a brief overview over the type of simulation modelling being used.
Marks will be given on:
(i) if the delivery of the talk is according to a standard that it could be used in teaching/presented at a
conference, (ii) your comprehension of the paper and how well you answer questions, (iii) if your slides
meet professional standards, and (iv) the general organisation of the talk and how well you covered the
aspects mentioned above. In the assessment, the criteria (i)-(iv) will have equal weight.
Talks will be scheduled after Easter, so have your talk ready at the end of Easter break, it can be
scheduled in any lecture/seminar slot after that date.
Coursework Assignment II (worth 70%)
A modelling problem is described in the first set of lecture slides (see material copied from the lecture
slides below). You are to address this problem using two different modelling techniques: first, using a
differential equation-based approach; second, using agent based modelling. Compare your findings
from both approaches and write a 6 page conference paper that summarises your findings. The 70%
marks you can obtain for this part of the coursework are split as follows:
Quality of the writing and figures in the report 10%
Technical work:
Develop a model based on differential equations addressing the problem;
that is, give a differential equation that models the problem, describe how
you derive it, and reason about its classification (7.5%).
40%
Explore the use of analytical techniques to gain insight into the system's
behaviour; that is, analyse the above differential equation, find full (or
equilibrium) solutions, and argue about their stability (7.5%).
Numerically integrate the differential equations using an appropriate
integration method of your choice and compare the results to the
analytical results; that is, develop the computer code for an integration
scheme (from scratch, do not use off the shelf libraries or integrated
functions provided by math software) and give evidence that it
reproduces system behaviour correctly, and reason about parameter
choices made in your integration scheme (10%).
Implement an agent-based model that addresses the same problem. Give
evidence (for example, appropriate example output) of the model you
have implemented, and compare your findings to the results obtained
with other methods (from above) (15%).
Use both models to answer the research questions given below.
Quality of an original extension of the problem:
Marked according to originality (Is this an interesting question to ask in
this context), quality of the analysis (do you understand what is going on
and did you use appropriate techniques to analyse the model), and its
motivation (can you convince me that it makes sense).
20%
Total for CW assignment II 70%
Reports are due at the end of term (May 15) and will have to be submitted in electronic form as a pdf
file. Don't forget to upload any simulation code you developed for the assessment.
Further instructions about the assessment can be found in the first set of lecture slides (copied into the
document here):
Problem description from lecture slides
Consider the following situation:
● A population of students is working on group projects. Students can follow two strategies (S): work
hard for the project (S=H) or free-ride (S=L).
● In every course, groups of size n are formed at random. Students use the strategy determined at the
beginning of the course (i.e. S=H or S=L) in their group work.
● Total group effort is determined by the composition of the group. In a group with h hard workers and
l=n-h lazy workers total group effort is e=h*H+l*L (H and L being the effort put in by hard/lazy
workers).
● When group projects are marked, every student in a group gets the same mark. The lecturer
determines this mark as m=e/n (that is, by dividing total group effort by the number of students in a
group; the larger this number the better the mark.)
● At the end of the semester groups are dissolved and every student rethinks his strategies for the next
semester. He/she does this by selecting another student (a reference student) at random and comparing
a measure ? based on marks and effort, =m-a*S (where a is a parameter and S=H or S=L depending
on strategy). The measure accounts for the mark obtained, but is lowered by the amount of effort spent.
Very good marks without effort maximise the performance measure. If the reference student selected
for comparison got a higher performance measure, the selected student will imitate the reference
student's strategy in the next semester with a probability that is proportional to the difference in the
performance measure. Students will not imitate strategies from other students with worse
performance measures.
● Students study forever (that is, take an infinite number of courses) and follow the same procedure (as
outlined above) for every course they take.
Research questions to address:
● Assuming we start with equal numbers of hard working and lazy workers, what is the composition of
the group
– After 4 years (i.e. 8 courses) if H=1 and L=0 and a=0.5.
– In the long run (after an “infinite” number of years)?
– How quickly is this equilibrium state reached?
● How do the following parameters influence results:
– Initial composition of the population
– Group size (n)
– Cost of effort, a
– Contribution of hard workers to group effort (i.e. H and L)