TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Junyan Chen Information System and Management studies 2003/2004

The candidate confirms that the work submitted is their own and the appropriate credit has been given where reference has been made to the work of others. I understand that failure to attribute material which is obtained from another source may be considered as plagiarism.

(Signature of student)

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Summary The role of taught modules timetabling is becoming increasingly important in the University. Although staff in many universities have tried to develop methods to help solve this problem, yet there does not seem to be a clear distinct timetabling method that has been used to help solve the problem in the University of Leeds.

The aim of this project is to deal with the problem of timetable clashes that students face when choosing modules for oncoming semesters, and try to develop some new algorithms or methods to solve this problem.

I

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Acknowledgement I would like to thank the following for their continued help, support and guidance throughout the various stages of this project.

Dr Kevin McEvoy

My project supervisor, who helped me decide which project to undertake and for his continued support and suggestions throughout the project.

Dr RSK Kwan

My project assessor, for his suggestions during the progress meeting.

Mrs Nicola Woolley

For her help in my understanding of the current timetabling system in Leeds Business School

Dr Clive Souter

For his help in my understanding of the current timetabling system in Joint Honours Department of Leeds University.

II

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Table of Contents Summary………………………………………………………………………………………I

Acknowledgement…………………………………………………………………………....II

Chapter 1: INTRODUCTION ………………………………………………………………1 1.1

Project aim…………………………………………………………………....1

1.2

Objectives, minimum requirements and deliverables…………………..…….1

1.3

Lecture timetabling…………………………………………………………..2

1.4

1.3.1

Definition…………………………………………………………….2

1.3.2

Why is timetabling difficult?...............................................................2

Traditional timetabling methods……………………………………………..3

Chapter 2: CURRENT TIMETABLING INFORMATION OF THE UNIVERSITY OF LEEDS ……………………………………………………………………………4 2.1

Current timetabling system ………………………………………………….4

2.2

Current Method of constructing timetable on School of Computing………..5 2.2.1

Example……………………………………………………………..7

2.3

Current Method of constructing timetable on Business School……………..8

2.4

Current method of constructing timetable of JHSC department…..……..….9 2.4.1

Set timetable in advance or afterwards?…………………………...10

2.4.2

Problems with current University system…………………..……...11

Chapter 3: RECOMMENDED METHODS AND ALGORITHMS………………..……13 3.1

Timetable Zoning………………………………………………..……..…...13 3.1.1

What is zoning………………………………………………..…….13

3.1.2

Why do we need to consider zoning…………………………..……14

3.1.3

Advantages of zoning………………………….………………..….15

3.1.4

Disadvantages of zoning…………………………………………...15

3.1.5

How many zones…………………………………………………...15

3.1.6

How many sessions…………………………………………….…..16

3.1.7

How have we allocated subjects to zones……………………….....16

3.1.8

Can some subjects be offered in more than 1 zone thus providing more combinations?....................................................................……16

3.1.9 3.2

Example………………………………………………………….….16

Sequential Method…………………………………………………..……...19 3.2.1

Timetabling using Graph Colouring………..……………………....20

III

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

3.3

3.2.2

Example 1… … … … … … … … … … … … … … … … … … … ..… .… ...20

3.2.3

Example 2… … … … … … … … … … … … … … … … … … … .… … .....22

Timetabling Programs… … … … … … … … … … … … … … … … … … … ..… .24

Chapter 4: Evaluation… … … … … … … … … … … … … … … … … … … … … … … … … … … 25 4.1

Summary… … … … … … … … … … … … … … … … … … … … … … … … … … 25

4.2

Evaluating the system against the project’s aims, minimum requirements and objectives… … … … … … … … … … … … … … … … … … … … … … … … … ...25

4.3

Evaluating the development methods and algorithms… … … … … … … … … 26

4.4

Future Enhancements… … … … … … … … … … … … … … … … … … … … … .26

Chapter 5: Conclusion… … … … … … … … … … … … … … … … … … … … … … … … … … ...28

References… … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … ...29

Appendix Appendix A… … … … … … … … … … … … … … … … … … … … … … … … … … … … 31 Appendix B… … … … … … … … … … … … … … … … … … … … … … … … … … … … 32 Appendix C… … … … … … … … … … … … … … … … … … … … … … … … … … … … 33 Appendix D… … … … … … … … … … … … … … … … … … … … … … … … … … … … 34 Appendix E… … … … … … … … … … … … … … … … … … … … … … … … … … … … 35 Appendix F… … … … … … … … … … … … … … … … … … … … … … … … … … … … 36 Appendix G… … … … … … … … … … … … … … … … … … … … … … … … … … … … 37 Appendix H… … … … … … … … … … … … … … … … … … … … … … … … … … … … 38

IV

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

CHAPTER 1 – INTRODUCTION 1.1 Project Aim This project will analyze the current system for timetabling taught modules at the University of Leeds. Then attempt to solve the existing problems, making some suggestion to improve the current timetabling system. Finally, this system will be easy for the students and staff to identify any timetable clashes and take appropriate action.

1.2 Objectives, Minimum Requirements and Deliverables The objects of this project are: 12 Understand and document the current system for timetabling modules at the University of Leeds 12 Investigate how the current timetabling system helps students to choose modules 12 Investigate the deficiencies of the current timetabling system

12 Investigate how to deal with timetable clashes

12 Improve timetabling system and avoid the class clashes 12 Develop a solution to solve the problem

The minimum requirements of this project are: 12 Description of timetabling system in the School of Computing 12 Description of the Business School timetable

12 Point of view of students and staffs from Joint Honours Department

12 Recommendations for improving the system The deliverable of this project is: 12 A project report Possible further enhancement for this project is: 12 Validating of timetable schedule

1

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

1.3 Lecture timetabling Organizations such as schools, colleges and universities schedule events (lectures, classes etc.) given limited resources (teachers, rooms etc.) and a pre-defined set of restrictions on the way resources may be exploited. Such restrictions range from the obvious (‘a person cannot be in two places at once’) to the more subtle (‘students should have a reasonable amount of preparation time between lectures belonging to the same course’).

1.3.1 Definition Lecture timetabling has been defined by [Ross 1994a] as follows: “Lecture timetabling is the problem of assigning times and places to many separate lectures, tutorials etc. to satisfy several constraints concerning capacities and locations of available rooms, free-time needs and other such considerations for lecturers, and relationships between particular courses.” University courses typically consist of a series of interactions between module staff and students at predefined times and places. Each such interaction is thought of as an event, and the participants as agents.

A timetabling problem can also be defined more formally (after [Ross 1994b]): 12 Given the sets E = {e1, e2… ev} of events, T = {t1, t2… ts} of times, P = {p1, p2… pm} of places and A = {a1, a2… an} of ‘agents’ (people whose presence is required for the event to take place; in the case of a University timetable for example, students and lecturers would be agents). 12 An assignment is a 4-tuple (e, t, p, a) such that,

12 A timetable is a collection of assignments, one per event.

12 The timetabling problem is now defined as the problem of finding a collection of assignments that satisfies or minimally violates a collection of constraints.

1.3.2 Why is timetabling difficult? There are an enormous number of ways in which events, times, places and people can be combined for even quite a small timetable. Each such combination will have its own advantages and disadvantages but (typically) only a few will satisfy all constraints and represent optimal solutions. The need for powerful techniques for solving timetabling can be appreciated by considering the fact that events can be scheduled in t time-place slots in et different ways. Finding optimal solutions can therefore prove extremely challenging.

2

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

1.4 Traditional timetabling methods A graph colouring algorithm is often used to solve timetabling problems either manually or automatically. Each event is represented by a vertex in a graph and edges are drawn between events which cannot occur concurrently. Colours represent time-place pairs. The algorithm can then be described as follows: 12 A vertex is chosen at random and assigned a colour corresponding to a time-place pair. 12 Every vertex connected to it is assigned a colour such that no joined vertices have the same colour. 12 If at any point it is impossible to continue (i.e. no colour can be chosen to colour a vertex in a way that does not break the restriction above) the algorithm is restarted.

One of the drawbacks of this type of algorithm is that it tries to find a feasible timetable rather than a good one: that is it does not take into account soft constraints. [Fang 1994] notes: “Conventional computer-based timetabling methods concern themselves with simply finding the shortest timetable that satisfies all the constraints, usually using a graph colouring algorithm and less with optimising over a collection of soft constraints. (Commercial software tools) … are usually based on straight forward constraint satisfaction techniques.”

3

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

CHAPTER 2 – CURRENT TIMETABLING INFORMATION OF THE UNIVERSITY OF LEEDS 2.1 Current timetabling system At the start of each semester, some students have to face a problem that it is impossible for them to take all modules (except compulsory modules) that they have chosen due to timetable clashes. Compulsory modules are modules students have to take. However, choosing optional and elective modules is only a worrisome problem to student, but also to the teaching staff. Students have to choose the optional module from their teaching departments. For example, a student doing “Information system” degree has to choose his/her optional module(s) from School of Computing. Whereas, in the case of, a student doing “Information system and management studies” degree, is a little more complicated. He/she has to choose the optional module(s) from two departments, which are “School of Computing” and “Business School”. Choosing the elective modules will be the most worrisome and complicated process. Students are free to choose modules from most of departments. Joint honours department students would face a greater risk than other students since by default they take modules offered by more than one department but any student that decides to take a module outside of their parent department are also at risk. The reason is because different department might result in timetables clashes. In fact, this situation occurs every academic year. So, some students are then forced to drop some of those modules they have chosen.

Students can check for timetables clashes from the “Central Teaching Space Timetable” web pages [SoC 2003b] or from the department module handbook. These web pages contain information for lectures’ timetable and location for all university modules. To find out the module detail you need, you have to select the department offering the module you are interested in. Personal experience states that it is quite useful when I chose my optional and elective modules. It makes the module detail more concentrated and easier to access. (See appendix B)

The process of constructing taught modules timetable is complex and requires a lot of experience. The staff to construct the timetable will gather a lot of information from different aspects. They might obtain from the past year experience, or from the current information they have got. A sophisticate staff will check the past taught modules timetable before they construct a new one. If the past timetable works, most of their work and time will be saved to construct the new timetable. For checking the changes between the past taught modules

4

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS timetable and the prospective ones, the staff will compare the taught modules in these two academic years. The staff will keep the modules which will still be taught in next academic year in the timetable, and remove those will not be taught in the next academic year. In fact, the changes will be tiny. Most of the taught modules will not be changed; only some of them will be removed or replaced. Therefore, the new modules will replace the blank time period which belonged to the cancelled or will be put into a more suitable time period.

2.2 Current Method of constructing timetable for School of Computing Currently, the School of Computing (SoC) uses a system called School Information System (SIS) to handle all network transactions within the SoC. A Postgres database is maintained within this system, which stores information regarding each student studying within the SoC. This database maintains a table containing timetable information for each module taught within the SoC. Other tables containing timetable slots and student information are linked to this table to enable a student or a staff of SoC to view their weekly timetable and information they need. This timetable information is available to the student via the Taught Students Office and the on-line SIS (School Information System) system. The on-line timetable has a server side dynamic link that connects the web pages via CGI (common gateway interface) scripts to call information from the Posrgres database. However, this system only stores and displays the timetable, not for designing.

Banner is the University of Leeds Student Information Management System. It is a large Oracle database that acts as the University’ s central records system and therefore holds data on all of its students. It is involved several of the Information Systems in the School of Computing.

Moreover, at the end of an academic year, SoC department confirms the provisional timetable for the next academic year. And, students would be required to provisionally to fill a form for choosing modules for the next semester before this semester has finished. However, the timetable details are not finalized; they could be changed when the new academic year starts.

In addition, the lecture can not be scheduled without room, and, it seems that it is impossible to go along the lecture if most of students are not free. Therefore, to schedule a lecture, at least three conditions have to be fulfilled:

5

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS 12 Picking time when students are free

12 Lecture rooms are free

12 Who takes the modules, i.e, the lecturer is free During the same period that the timetable information is confirmed, students studying within the SoC are required to provisionally select modules they wish to study in their forthcoming year. The current method of selection is a paper based form issued by the SoC and each student must select modules by completing this form.

[Cranshaw 2002] The only aid a student has when completing this form is a handout (see appendix B) made available by the Taught Students Office in the SoC. This handout details SoC module timetable information for the next academic year. Students select modules they are interested in by referring to the SoC Student Handbook, before making their final decisions whilst considering credit, level and prerequisite requirements of each module.

In retrospect this appears straightforward but problems arise for students when trying to construct a list of modules they wish to study without any timetable clashes. It can be very frustrating when choosing modules that they are interested in, because the checking for clashing modules has to be done by hand using the handout. This frustration increases further for students taking module electives outside of the SoC because they are forced to cross check the SoC paper based timetable with the module elective handbook manually. The process is even more uncertain for joint honours students who have to compare at least one of the two sources already mentioned with their other joint honours department’ s internal timetable.

This entire process it very testing for all SoC students who firstly have to spend the time finding the modules they are interested in and then are forced to cross check up to three sources of information in order to work out if their preferred modules clash. If they then find a clash, they have to start the process all over again until they have found enough modules that they are interested in which do not clash. The frustration is compounded by the fact that although live in a world of Information Systems, students are forced to consult up to three sources of information to create a paper based timetable, rather than just one. Students are effectively examining three data sources from the same Information System as the UoL holds all timetable information on modules taught at the UoL (see figure 1).

6

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Decide on prospective modules

UoL Central Timetable Information

Check

Check

SoC timetable information

Check

Module elective handbook timetable information

Other departments timetable information within the UoL

timetable information

clashes found

Construct manual timetable

No clashes found/ module timetable changed

Process complete

Figure 1: Flow diagram of student activities when deciding upon modules using the current system.

2.2.1 EXAMPLE Here is the time slots example for selecting the time period(academic year 2003/2004, semester 1) COMP1620 (GI11) Introduction to Human Computer Interaction

Class Size: 187 Thursday

10:00:00

Friday

13:00:00

The numbers indicate the numbers of students who are already timetabled in that slot in the University Timetable. Note that students may not be required to attend at all timetables slots (eg. Duplicate lectures or heavily timetabled Chemistry practicals). Because of the existence

7

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS of duplicate lectures a student may already have an apparent clash, and will be counted twice in such a slot.

9am 10am 11am Noon 1pm 2pm 3pm 4pm 5pm

Monday 24 2 49 31 185 135 47 148 0

Tuesday 3 3 3 137 387 156 62 153 23

Wednesday 130 129 161 4 10 1 0 0 0

Thursday 26 187 1* 29 153 62 138 127 23

Friday 23 0** 58 19 194** 185 176 2 0

*table 1: of all the students on COMP1620 only one student has a commitment on the university timetable at 11 am on Thursday. ie. One of that student’ s modules has a lecture or class at 11 am Thursday. ** As two students doing Joint Honours with physic have a clash with COMP1620, the lecture in 1pm had to be moved to 10am on Friday. And, as 2 modules have been scheduled at the same time, seven students have been duplicated counted.

2.3 Current Method of constructing timetable on Business School The reasons that I choose Leeds University Business taught modules timetabling system for researching are, firstly, most of the students in Joint Honours Department are taking the JHSc (Joint Honours in Science) degree, and, many of these students have a same teaching department – Business School. Secondly, there are quite a lot modules can be taken in Business School, and some of these modules are 2-hour slot, which would be the special and difficult cases for arranging the timetable. Thirdly, several modules in Business School have a very small number of students to be taken, which would provide a special case for the project.

Compared with the SoC, the timetabling process of the Leeds University Business School is quite similar to the SoC. However, they are more relying on the paper-based work. An interview making with Business School Undergraduate Office Secretary, Mrs Nicky Woolley could help me to fairly understand how they arrange the timetable. As the other department staff does, Business School staff has to distinguish all the compulsory modules and optional modules, to make sure the compulsory modules for all student taken will not clash with others. Before they try to arrange the timetable, they will consult the last year’ s timetable first. They would suppose all modules timetable would not be changed. If some modules are modified or cancelled, they will fill those blank modules’ period with the new ones. Furthermore, the staff have to consider some the other factors is arranging the timetable, such

8

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS as, which semester the module will be taken, which group of students will take, how long the lecture will take, ie, one-hour slot or two-hour slot, how many students are taking, what equipments required for the lecture, and so forth. Nicky told me that some very special and “ headachy” situations exist in Leeds University Business School, one of which is most of the taught modules have a great size of student population. As the large lecture theatres are very limited in Business School, even the entire University, it is very difficult to arrange a suitable timetable to be fulfilled. When some Business School modules got clashes with other departments, then the programme staffs have to negotiate with that department to arrange the suitable time, or try to duplicate the lectures, and which will increase teaching cost as well. In addition, when a two-hour-slot module in Business School clashes with another taught module, no matter which department it is, the Business School programme staffs will try to persuade that department to change the module timetable, rather than the Business School does. Once that two-hour-slot module is changed, it would be more difficult to get a suitable time period, more students will be influenced, and more clashes would happen.

2.4 Current method of constructing timetable of JHSC department It seems to be a small case replacing a taught module by another one in the same department. In fact it is not, even, it could cause a big change in the timetable. This phenomenon would be more outstanding in JHSC department.

A meeting with JHSC department director Dr Clive Souter help me to understand the current timetabling system in Joint Honours Department. According to the introduction by Dr Souter, I realized that JHSC department would not be responsible to schedule the taught module timetable. The timetable of the JHSC student is arranged by their teaching departments. The module timetable of a JHSC student who is taking a Information system and Management studies, for instance, is scheduled by School of Computing and Business School, rather than JHSC department. However, staffs of the JHSC department are able to check JHSC students’ timetable if they want to. They can check the timetable using the software called Splus produced by SCIENTIA Company.

In addition, it would not be a big problem if the new module does not clash with other department’ s taught modules. The staff just refill the replaced module time period would be all right. However, if new module clash with other department’ s taught modules, there are more steps to be done. Supposed this new module is compulsory for Joint Honours Information system. Firstly, the staff have to know all the Joint Honours degrees with

9

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Information System (all the Joint Honours degrees with Information System have been list in a table below). As a compulsory module for information system, it must not clash with other half part of the Joint Honours degree. For example, Object Oriented Analysis and Design (COMP2470), a compulsory module for Information system in Level 2. It must clash with neither any of Accounting compulsory modules nor Biology. Or clash with neither Management compulsory modules nor Computing compulsory modules, and so on. And these compulsory modules are supposed not in the same time period.

If the clash happens, departments have to negotiate a suitable time to arrange the timetable with each other. The way to solve the timetable clashes is duplicating this module lecture or dividing the student into different group. And these lecture or groups will be help in different time, which depends on the students which one they preferred to go. This way is usually adopted by Business School of Leeds University, as many of the Business School modules have a large number of students to be taken. Timetable clashing seems can not be avoided. Of course, when staffs want to use this method to solve the timetable clash, they have to several aspects: 12 The amount of students taking that module. As only a few students taking that module, duplicating the lecture will either increase the teaching cost of the university or waste the resources. 12 How many students clash? I don’ t think it is necessary to duplicate a module lecture, as only one or two student clashes the timetable. 12 Is there an alternative? It means that is it the only and best way to solve the problem?

2.4.1 Set timetable in advance or afterwards? There is a debate on whether to set the taught timetable before the students choosing the optional modules for next semester or afterwards. As I mentioned before, the students have to choose the modules for next semester before this semester finished. A draft timetable for the next semester is necessary, as it could help them to know the clashes that might happen. However, as this draft timetable is based on the past information, the information might be changed in the next semester. The past information would include the past taught module timetable, past prospectus or even their experience. The popular modules for last year might get less attraction to the students. In contrast, the less popular modules might attract more students in next semester. Therefore, it is not good for the staff to arrange the timetable.

10

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS In my opinions, I would support setting the taught timetable afterwards. As setting the timetable afterwards would help the staff to arrange the taught timetable more easily. After the students choosing the optional modules for the next semester, staffs can get the most updated information from this year. Information will be more veracious and updated. According to the situation of this year, timetable can be set well and truly. But the disadvantages might be: the students would have a misconception which they think they can choose any optional modules without clashing, and which would make a bad sense to students for choosing the optional modules

The project instigator Dr. Kevin McEvoy had provided a copy of a Microsoft Access database that contained class times and location for all modules taught within the University of Leeds. The data in Microsoft Access was supplied to Dr. McEvoy form the University’ s ISS (Information System Services) the University’ s “ Central Information Technology and Information Provider” . Moreover, I was told that there is a shortage of the timetable of a third year module COMP 3670 (SY32) “Secure computing”. It is a second semester SoC modules. Its lecture time and location have not been confirmed. Students can not get enough information to choose their optional modules for the next semester. It might cause a mess for the next semester timetable arrangement. Where is SY32? It should be somewhere in the table below. This table is available from the web site of SoC during semester 1 (See appendix C). Eventually, there still got one student had a clash, and that student had to drop that module which is an optional one to him.

2.4.2 Problems with current University System Although there are a few software and systems to help for the timetable arrangement, according to the research, however, timetable clashes still can not be avoided. The major reason for accruing this is because each department doesn’t have information from others. i.e. lacking of communication. Each department set up their timetable individually; they don’ t have the other department’ s timetable. So, the first thing need to improve is the communication between different departments. In my opinions, the entire university timetable should be arranged by a single department. For example, Joint Honours Department is a very special department which has a potential relationship with other single departments. It should have all modules information from different departments, as staffs need to use that information to deal with their job. So, one of the advantages of arranging the entire university timetable in a single department is able to minimise the mistake from the communication. Another advantage would be able to minimise the human errors from the

11

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS process of timetable arrangement. As fewer staff involving in this process, the fewer human errors will arise. Those staffs can get concentration on the process of timetable arrangement. That would make it better and better.

In fact, it is not so easy to find out whether the clashes occur or not. In most of the cases, clashes are reported by the students, then the staff to confirm whether it happened or not. If the clashes occur, then staffs have to arrange that module to another suitable time period.

12

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

CHAPTER 3 – RECOMMENDED METHODS AND ALGORITHMS This chapter provides several methods to solve the taught modules timetable clashing problem.

3.1 Timetable Zoning I read the information from the web sites of University of Wolverhampton. They advocate an effective and logical method called “Timetable zoning” [Internet 03] for timetabling taught modules. In fact, “Timetable zoning” method has been trailed in some universities, such as, University of Wolverhampton. The fact proves that this method has gained a great success in timetabling taught modules for the university. Therefore, I would like to introduce “Timetable zoning” method and analysis whether it is suitable for the timetabling taught modules system of Leeds University or not.

The reason of why “Timetable zoning” is so useful as it ensures that students doing the joint or combined degrees are able to access their core modules without academic and registry staff having to go through each programme, which also can save the time and reduce the human errors. The principle of zoning is to split the teaching week into certain zones. If those zones certain combinations of subjects will clash, then, each zone will be divided into certain subzones. The implication of this is that subjects in combined and joint degrees will not be able to designate more than 2 core modules at each level and will be encouraged to limit core to less than this. The number of core options and where they appear on the timetable is however not restricted.

3.1.1 What is zoning Take the example of Level 1 taught modules timetables in Table 3 (details from Appendix D): Zone1 might be: Monday

Tuesday

Wednesday

Thursday

9:00-11:00

12:00-2:00

9:00-11:00

3:00-5:00

Zone 2 might be: Monday

Tuesday

Friday

Thursday

3:00-5:00

9:00-11:00

9:00-11:00

12:00-2:00

Zone 3 might be: Monday

Tuesday

Thursday

Friday

12:00-2:00

3:00-5:00

9:00-11:00

12:00-2:00

13

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS As Table 3 shows that, the time periods within the zones do not overlap so that any subject allocated to one zone is guaranteed to be able to be taken with another subject provided it is in a different zone. Moreover, Table 3 also shows how this could operate with 3 zones, 3 sessions per day and 3 undergraduate levels of provision. An instance of arranging modules timetable of a student who is taking a Joint honours degree in Information system and management studies could be like this: computing modules can be arranged on any of those three zones in any session first; then, you can choose one of the rest of two zones in any session for the management modules. No time period will overlap.

“ Timetable zoning” is designed to solve the problems of subject, rather than solving the problem of timetabling specialist or named degrees. However, in timetabling such degrees,

where it is intended to use the same modules as those used in combined degrees, the place of these shared modules will have to be within the appropriate zone if the module is only being offered once i.e. if there is only one iteration of the module.

3.1.2 Why do we need to consider zoning The basis for the modularity project [Internet 03] is that students should make choices at the programme level and much less at the module level. This would reduce the amount of academic input needed to validate programmes and should in principle allow swift module registration and hence, given better information about module populations, enable better timetabling. It is also inline with the need for programme specification.

The current approach is to (a) Take the existing module offerings in each subject in their current timetable slot (b) Attempt to predict the most popular subject combinations (c) Attempt to timetable each combination

This leads to great inefficiency and is already proving to be an almost impossible task for the subject leaders who are being asked to negotiate with up to 20 other subjects.

Promoted combinations selected for academic reasons could be timetabled by the present system (with considerable effort) but this relies on using alternative modules and/or duplicating modules from each subject in order to satisfy the timetable constraints. If a subject is revalidated or adds/removes modules the whole process of re-negotiation for this subject will need to be done again. Many modern universities have adopted a form of zoned

14

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS timetables in order to rationalise part of this process and whilst this approach will only solve part of the problems it does present a means forward if only to solve the timetabling of combined (subject) and joint honours degrees.

3.1.3 Advantages of zoning 1. Guarantee to be able to timetable a student programme based on a combination of subjects which are in different zones. This facilitates joint honours and combined degrees. 2. Subjects have flexibility to arrange modules within a zone and if the portfolio changes or the sessions change the guarantee in (1) above still stands 3. It is possible to move modules within a zone in order to cater for teaching room shortages, staff availability etc. 4. Academic development is not hindered by the fact that any changes would not have to be negotiated with other subjects to ensure that student' s programmes are still valid and able to be timetabled. 5. There is no limit to the options that can be provided but in order to guarantee availability they must be within the allocated zone i.e. the timetable is devolved to the subjects but within closely defined rules.

3.1.4 Disadvantages of zoning 1. Not all subject combinations are possible i.e. subjects which are in the same zone cannot be combined 2. Subjects will be limited as to how many modules can be offered at each level options can be provided but given the number of sessions available these may be restricted 3. Modules must be associated with one subject only or repeated There are ways to overcome some of these restrictions but the principles will always hold.

3.1.5 How many zones [Internet 03] says that the more zones the greater the number of subject combinations but each subject will then have to fit into a smaller number of sessions which restricts the number of modules on offer at each level. If we assume 13 sessions i.e. not Wednesday pm or Fri eve then 3 zones allows 4 sessions per subject and 4 zones allows 3 sessions for a subject. If all

15

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS modules fitted into a session then 4 zones may be feasible but for band 2 subjects this is unlikely. Three zones this seems to be the norm for other universities who have used this approach and we have adopted this number.

3.1.6 How many sessions It has been assumed that we work with 3 sessions per day, each of 3 hours. With zoning there is no reason why we should not use 2 or even 1-hour sessions!! However the 3-hour session is often the most appropriate for part-time students.

3.1.7 How have we allocated subjects to zones It is possible to optimise this if we can prioritise which combinations are the most important. The University of Leeds has taken the combinations which are currently in UCAS guide and weighted these according to the numbers of students who have taken these combinations. With 3 zones it has been possible to timetable 83% of the proposed combinations.

3.1.8 Can some subjects be offered in more than 1 zone thus providing more combinations? In many ways this has already happened e.g. Business and Computing. Computing is likely to offer in one zone only but a variation Information Technology will be offered in more than one zone. By offering a subject in 2 zones it can be combined with any subject offered. There are several subjects in the same category as Computing and Business e.g. Psychology, English, History

3.1.9 Example Try to apply the “ Timetable Zoning” method to arrange the timetable for some taught modules of Leeds University.

I chose some programs of JHSc department of Leeds University for this example, which are: BSc Joint Honours degrees (SoC modules join with other department modules): 12 Computing and Accounting

12 Information Systems and Accounting

16

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS 12 Artificial Intelligence and Mathematics 12 Artificial Intelligence and Physics

12 Computer Science and Economics

12 Computer Science and Mathematics 12 Computer Science and Music

12 Computer Science and Physics 12 Computing and French

12 Computing and Management 12 Computing and Philosophy

12 Information Systems and Management Studies And a SoC single honours degree: 12 BSc Honours Cognitive Science (Modules details: see Appendix E & F)

For this example, I chose the compulsory modules of these degrees for arranging the timetable. Other optional and elective modules will not be taken into account.

The process of arranging the timetable will be shown via three tables which are table 4a, table 4b and table 4c.

Table 4a is obtained from Appendix D, which has been divided one teaching week into four zones. And I have coloured these four zones into 4 different colours: Zone 1 = yellow, Zone 2 = light blue, Zone 3 = purple At the beginning, the School of Computing subjects were allocated into zone 1, the Business School subjects are in zone 2, and the rest of the title were allocated in the zone 3. According to the different situation those subjects engaged a single or multiple sessions. For example, computing subjects were in the session of Monday 12:00 – 2:00, and Economic engaged two sessions which are Wednesday 9:00-11:00, and Thursday 3:00-5:00.

17

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Table 4a

Level 2 Monday Tuesday Wednesday Thursday Friday

9:00-11:00

12:00-2:00

3:00-5:00

Level2-Zone2

Level2-Zone1

Level2-Zone3

Level2-Zone3

Level2-Zone2

Level2-Zone1

Level2-Zone2

Free

Free

Level2-Zone1

Level2-Zone3

Level2-Zone2

Level2-Zone3

Level2-Zone1

Free

Table 4b

Level 2 Monday Tuesday Wednesday Thursday Friday

9:00-11:00

12:00-2:00

3:00-5:00

Accounting

Computing

Physics

Mathematics

Management

Information system

Economics

Free

Free

Computing Science

French

Economics

Philosophy

Artificial Intelligence

Free

Specifying the module time period would be shown in the following table:

Table 4c:

Level 2

Monday

Tuesday

Wednesday

Thursday

Friday

9:00-11:00

12:00-2:00

3:00-5:00

Accounting: LUBS2200(9:00-11:00) Management: LUBS2235 (11:00-12:00)

Computing: COMP2400(12:00-1:00) COMP2600(1:00-2:00) COMP2400(2:00-3:00) Psychology: PSYC2040(12:00-2:00) PSYC2011(2:00-3:00)

Physics: PHYS2042(3:00-5:00)

Management: LUBS2235(12:00-1:00) LUBS2195(1:00-3:00)

Information system: COMP2470(3:00-4:00) COMP2470(4:00-5:00) Accounting: LUBS2220(3:00-5:00)

Free

Free

French: FREN2011(12:00-1:00) FREN2011(1:00-2:00) Mathematics: MATH2080(2:00-3:00)

Economics: LUBS2280(3:00-4:00) LUBS2340(4:00-5:00) LUBS2240(5:00-6:00)

Artificial Intelligence: COMP2340(12:00-1:00) COMP2340(1:00-2:00) Management: TRAN2010(12:00-1:00) TRAN2010(2:00-3:00)

Free

Mathematics: MATH2420(9:00-10:00) MATH2080(10:00-11:00) MATH2420(11:00-12:00) Computing: COMP2600(9:00-10:00) Economics: LUBS2040(9:00-10:00) LUBS2050(10:00-11:00) LUBS2280(11:00-12:00) Computing Science: COMP2360(9:00-10:00) Management: LUBS2235(9:00-10:00) LUBS2850(10:00-11:00) LUBS2240(11:00-12:00) Accounting: LUSB2210(9:00-11:00) Philosophy: PHIL2020(11:00-12:00)

18

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS As three subjects, which are Music, Philosophy and Geography, are all optional modules in Level 2 semester 1, they would not be taken into account in this example. More details are shown in Table 4b.

The next step to arrange the timetable will concentrate on how to arrange the specific modules into the different sessions at a suitable time period. As different modules have different time slots, some may be 1-hour, and some may be 2-hour. Moreover, the amount of the lectures taken should be taken into account. As some modules may have 1 lecture in a teaching week, and some might have 2. Therefore, we should recognize these two issues for each single module as well. More modules detail, see Appendix G.

In this example, all School of Computing modules are 1-hour slot and held 2 lectures in one week. In total, COMP2400 and COMP2600 require 4 hours in a week. The session on Monday 12:00-2:00 has not enough time to be allocated, therefore, one of the lectures will be move to another session, and for example, one of COMP2600 lecture was set in zone 3 from 9:00-10:00. The reason I put this computing module in this session as there is no combination degree between Computing and Mathematics. Even though the time clashes, it would not affect the students. In the same way, I set the specific lecture time for the other 3 sessions (Tuesday 3:00-5:00, Thursday 9:00-11:00 and Friday 12:00-2:00) in Zone 1(see table 4c yellow zones). Zone 2 (light blue zone) consists of Business School modules. As those three Accounting modules are 2-hour slot, I out them into three different sessions, which are Monday 9:00, Tuesday 3:00 and Friday 9:00. Economic modules engaged two sessions in zone 2 as the large numbers of lectures.

The result of using “ Timetable Zoning” method to arrange some compulsory modules has been shown in Table 4c.

3.2 Sequential Method The following are some methods of solving timetabling problems taken from a similar project undertaken by the School of Computer Science and IT at the University of Nottingham.

One approach to the timetabling problem that I will discuss and evaluate is the sequential method. The basic principle behind the sequential methods is that they order events using domain heuristics and then assign the events sequentially into valid time periods so that no events are in conflict with each other. Of course, with different timetabling problems the

19

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS aforementioned events dependent on what type of timetabling problem is being dealt with. For instance, in modules timetabling the events would be the modules and the lectures, and, in exam timetabling the events would be the exams themselves, however, in the bus timetabling the events would be the journeys between bus stations.

3.2.1 Timetabling using Graph Colouring On the subject of timetabling, the problem in question could be represented as a graph with the events being symbolized by vertices of the graph. Conflicts between the events, i.e. violation of given constraints, would then be represented as edges, de Werra(1985). If there were an edge present between two nodes, this would signify that scheduling those two events at the same time period would violate constraints. For example, if we consider university lecture timetabling then if a student has to attend two lectures at one particular time, then there would be an edge between the two nodes corresponding to the two lectures in question. Therefore to produce a conflict-free timetable, or at least a timetable with all hard constraint adhered to, the sequential method of Graph Colouring can be used. The basic idea behind any particular graph colouring problem is to colour in the nodes of the graph in such a way that no two nodes that are connected directly via an edge have the same colour. If this were done, then the completed graph with all nodes coloured in would represent a feasible schedule for the given problem. If there were a restriction on the number of events scheduled at one particular time period then a rule would be put in place starting that only a certain number of events could be coloured in one particular colour.

3.2.2 Example 1 Below are the examples that will be used to help demonstration a standard timetabling problem. Information:Joint Honours in Science (JHSC) offers over 70 programmes of study covering a wide range of subjects. Now I just take some combination courses with information system or management studies in semester 1 of level 2, as an example from this project. In Leeds University, the courses combining with information system or management studies include: 12 Accounting- Information Systems 12 Geography- Information Systems

12 Information System- Management studies

12 Biology- Management studies

20

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS 12 Chemistry- Management studies

12 Computing- Management studies

12 Geography- Management Studies 12 Management-Mathematics 12 Management- Statistics

As there are two pathways for choosing the BSc Joint Honours in Management courses at level 2, we have take all of those compulsory modules into account for scheduling the timetable. The compulsory modules for some JHSC subjects have been listed in Appendix E. The following diagram shows the links among the modules: And the following compulsory modules have students studying both subjects and can not be scheduled at the same time period:Chemistry and Management

Mathematics and Management

Statistics and Management

Computing and Management

Information System and Management

Accounting and Management

The modules have each been given a code that is used in the graph: Mathematics = Math, Statistics = St, Management = Ma, Chemistry = Ch, Computing = Co, Accounting = Ac, Information System = IS

Diagram 1: the links among the modules

Ch CHEM2110 CHEM2190 CHEM2420

St MATH2710 MATH2730

Math

Ma

MATH2080 MATH2420

LUBS2235 LUBS2850 LUBS2195

Co COMP2400 COMP2600

Ac LUBS2200 LUBS2210 LUBS2220

IS COMP2470 COMP2400

21

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Therefore, the timetable for compulsory modules for Joint Honours degrees with information system or management studies could be like the following table (Table 4):

9 10

Monday

Tuesday

MATH2710(9am-10am) CHEM2190(9am-10am) LUBS2200(9am-11am)

CHEM2190 MATH2730

Wednesday

Thursday

Friday LUBS2235

MATH2080

MATH2730(9am-10am) COMP2600(9am-10am) CHEM2420(9am-11am)

MATH2710(11am-12pm) CHEM2420(10am-13pm)

11

LUBS2850

LUBS2220

12 1 2 3 4 5

CHEM2110

LUBS2235

LUBS2235

MATH2080 COMP2400 MATH2420

COMP2400 COMP2470

CHEM2420

MATH2420 LUBS2195 CHEM2190(4pm-5pm) LUBS2210(4pm-6pm) CHEM2110(5pm-6pm)

COMP2600 COMP2470

3.2.3 Example 2 This method can be used to arrange the exam timetable as well. Here is the example:

Information:10 exams

2 rooms

300 students

3 days with 2 slots per day

For this example we will have 300 students sitting either one or two exams and no more. Exam subject

Number of candidates

Maths

50

Statistics

30

English

40

History

30

Management

50

Computing

50

Physics

30

French

30

German

30

Chemistry

40

We suppose the following exams have students studying both subjects and can not be scheduled at the same time period:-

22

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Maths and Statistics

Maths and Management

Maths and Chemistry

Maths and Computing

Statistics and Management

English and History

English and French

Management and Computing

Physics and Chemistry

French and German

The exams have each been given a code that is used in the graph: Maths = M, Statistics = S, Management = Ma, Chemistry = Ch, Computing = Co, History =H, German = G, French = F, Physics = P, English = E.

Ma

Co

S

M

Ch

P

E

F

H

G

The completed graph above corresponds to the following exam timetable:Day Day 1

Day 2

Day 3

Room

Morning

Afternoon

1

Maths

Computing

2

Physics

Statistics

1

Management

English

2

Chemistry

German

1

French

2

History

23

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

3.3 Timetabling Programs Numbers of computer timetabling programs can be used for a vast array of timetabling problems. The main advantage of using timetabling programs over working out a solution by hand is that the speed at which computers can operate at means a solution can be achieved in a much shorter space of time. The diversity of timetabling problems means there is not always a suitable timetabling program to be used for the users specific problem. However, for lecture timetabling there is a wide variety of programs available such as Facility CMIS, Lantiv and aSc Timetables. Each individual timetabling problem is different, with different constraints that have to be adhered to, and so even when using a timetabling problem it is sometimes necessary to finish off a timetable schedule by hand. Computer timetabling problems are very good at satisfying some constraints, in that they can create a feasible solution if one exists. And, these constraints could be: 12 No entity must be demanded to be at more one place at a time. In taught modules timetabling this would mean no student can sit more than one lecture at any one time. 12 For each period in the timetable, the resource demands made by the events scheduled for that period must not exceed the resources available. In room timetabling this means it is not allowed to seat more students in a room than there are desks.

24

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Chapter 4 – Evaluation 4.1 Summary This chapter evaluates the complete project. The main areas of discussion will be evaluation the methodology employed and a project management analysis. I will also be reflecting on whether the project has achieved its aims, and how comprehensively they have been covered.

4.2 Evaluating the system against the project’s aims, minimum requirements and objectives The project has met all the initial requirements and each of the objectives set. The aim of this project was to develop a new methodology or algorithm to solve the existing problem of the Leeds University timetabling system. This was achieved by performing background research on how the current timetable is implemented:

To report of the current state of the module timetabling system of University of Leeds It is felt that this minimum requirement has been met. The current state of the module timetabling has been discussed and some of the major timetabling problems that are deliberated today have been identified.

To identify the different methods used in timetabling and to explore how they operate. Many methods of timetabling that are used in the world today have been examined, and explored in detail how they operate, and their advantages and disadvantages have also been discussed.

To examine which methods of timetabling are pre-eminent. Having identified many different methods of solving timetabling problems it became apparent that many of the methods are valuable in different situations, so to determine that one particular method was better than all the rest would be wrong.

To solve a small sized module-timetabling problem demonstrating the wider range of timetabling problems that can be used. A relatively small module-timetabling problem was constructed and I set about solving it using heuristics that were advantageous for me to solve the problem. The problem was then

25

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS solved using some of the timetabling methods that had been discussed to illustrate how the timetabling methods operated.

4.3 Evaluating the development methods and algorithms For the process of the algorithms itself the author has, somewhat, mixed views, two custom algorithms, which were “ Zoning Timetabling” and “ Sequential Method” , were used for this project. Both of them had been applied in the Leeds University Timetabling system, which could be traced back to Chapter 3 of this project. And, in my opinion, those two algorithms could work well for the Leeds University module timetabling system.

The methodology that was used was to demonstrate and discuss the diverse methods of module timetabling, using my own small module-timetabling problem to demonstrate some of these methods. It is felt that this was indeed achieved, although it would have been desired to demonstrate more of the timetabling methods that were discussed, but time restrictions prevented this. It was constructive to produce a schedule to enable the project to be completed by the deadline, and this involved imposing deadlines for different phases of the project. It was found that completion dates for the background reading, which was an important part of the project, and also the completion of understanding many of the different timetabling methods were found difficult to meet. However a period of time had been allotted to account for any such occurrence. Therefore overall it could be considered a successful in managing the available time over the whole period to get the project completed on time, but the actual different phases of the project were not completed within the self imposed completion dates. It is impossible to estimate how many hours of work were spent on this project.

4.4 Future Enhancements There is a very wide scope for future enhancements that could be incorporated into the developed algorithms, ranging from minor alterations to extension of the system. Only a

handful of timetabling methods have been discussed in this project, and so there are many more methods that can be examined in much greater detail if there is time. This project’ s primary objective was to identify some methods of timetabling and demonstrate how they operate. The notion of combining timetabling methods to create a more functional method has been touched upon. Depending on the timetabling problem, it could be constructive to combine different methods, and further work

26

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS could be employed to establish how a combination of certain timetabling methods might operate. Moreover, the combination of using the developed algorithms and the specified module timetabling software in the university timetabling system would have a bright future.

27

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Chapter 5 – Conclusion The aim of this project was to develop a new methodology to solve the existing problem of the Leeds University timetabling system. This was achieved by performing background research on how the current timetable is implemented. After the initial understanding of the how the existing system worked, I was able to investigate what problem the timetabling system existed. Once an understanding of what problems existing in the timetabling system, it could be easier to develop a new or better method to solve.

The development of this project was quite challenging because I had no experience of arranging the timetable in a university-wide scale, and I had little skills and knowledge on project management and module timetabling. The methodology used gave a solid base to work from and allowed me to plan the various stages effectively throughout the entire project process. I now feel much better equipped to engage on timetabling projects because of the knowledge and experience gained from completing this project. There were many problems I had suffered during the project’ s lifecycle. The major problem experienced was the method developing. Although I had got a lot of information for module timetabling, it was difficult to find out a better and easier one from the numerous of information. And when I had chosen the method, I had to apply it in the Leeds University timetabling system. Each of the tools considered had many benefits and weaknesses but, with hindsight, I believe the correct choices were made because the system has exceeded all expectations and works, as it should.

28

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

References 12[Atherton 2002]: Atherton P, (2002), Two subjects, too hard a choice, School of Computing, University of Leeds. 12[Cranshaw 2002]: Cranshaw J, (2002), Web based timetable builder for SoC students, School of Computing, University of Leeds. 12[Paul 2002]: Paul Cookson, (2002), Presentation and Application of the Methods of Timetabling, School of Computing, University of Leeds. 12[Cunningham 2003]: Cunningham S, (2003), Information System Services, University of Leeds, URL: http://www.leeds.ac.uk/iss/ [17th April 2004] 12[SoC 2003a]: School of Computing, (2003), SIS – the School Information System, URL: http://www.comp.leeds.ac.uk/internal/sis/ [10th April 2004] 12[SoC 2003b]: University of Leeds, (2003), Central Teaching Space Timetable, URL: http://www.leeds.ac.uk/timetable/ [20th April 2004] 12[Simon 2001]: Simon Bennett, John Skelton, Ken Lunn, (2001), Schaum’s Outlines UML, McGraw-Hill International (UK) Limited. 12[Internet 01]: URL: http://builder.com.com/5100-6388-1050671.html [14th April 2004] 12[Internet 02]: URL: http://www.dcs.warwick.ac.uk/~csvkn/index.html [14th April 2004] 12[Internet 03]: University of Wolverhampton, URL: http://www.wlv.ac.uk/project/timetable/overview.doc [20th April 2004] .

29

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS 12[Internet 04]: School of Computer Science and IT, University of Nottingham, (2002), Automated Scheduling, URL: http://www.cs.nott.ac.uk/~sxp/Scheduling/UnivTT.pdf [19th April 2004] 12[Internet 05]: School of Computer Science and IT, University of Nottingham, (2002), A Memetic Algorithm for University Exam Timetabling, URL: http://www.asap.cs.nott.ac.uk/publications/pdf/patat95_algorithm.pdf [18th April 2004] 12[Suresh 01]: Dr. Suresh Manandhar, Final year project: Timetabling using Cellular Genetic Algorithms with Adaptive Mutation Operators, Computer Science Department, University of York 12[Werra 1985]: De Werra D, (1985), An introduction to Timetabling, European Journal of Operational Research. 12[Ross 1994a]: P. Ross, D. Corne and H. Fang 1994 – Successful Lecture Timetabling Using Genetic Algorithms 12[Ross 1994b]: P. Ross, D. Corne and H. Fang 1994 – Improving Evolutionary Timetabling With Delta Evaluation and Directed Mutation 12[Fang 1994]: H. Fang 1994 – Genetic algorithms in timetabling and scheduling

30

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS

Appendix Appendix A: Personal Reflection Initially I was rather apprehensive about undertaking a project of this size. However, now looking back, I found the experience to be very beneficial and also satisfying. It was a chance to articulate my ideas and development strategies in a personal way, building a project to my own design, with only a basic guide structure being dictated. This is very different from what the average coursework allows.

There is indeed a lot of work involved in the project, and the task can look quite daunting from the outset. However I found it was very important to create a well thought out and flexible project plan. Spending time here at the start of the project saved me a great deal of time and effort later on. Having to revise the plan, in most cases, cannot be avoided; however, having a well-constructed plan initially greatly assisted the process. I found it made good sense to produce the project in stages, rather than writing the whole thing at the end. Sections on background research are best written when the material is fresh in ones mind, and method sections can be gradually added to as the work is carried out.

The project has improved my time management skills, and also my ability to plan and create a formal document, summarising findings and arguments.

This project has met all the requirements, objectives and deliverables set in its initial stages. I am very satisfied with the success of the project and I have certainly attained my personal objectives. These included enhancing my programming skills and experiencing the management of a project over a long period of time.

31

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix B: Level 1 teaching timetable of School of Computing

32

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix C: Level 3 teaching timetable of School of Computing

Monday 9 10 11 OR31 LT7 12

Timetable for Level 3 Semester 2 2003/2004 Tuesday Wednesday Thursday SC33 LT1

CS32 LT11 SY31 EERLT

AI32 ALL

AI34 PBB11

OR33 LT11

IS31 LT23

AI32 LT10

SE31 LT18

IS31 ALL

DB32 MELTA

1 SE31 LT16

AI34 PBB11 CS31 LT12

2 SY31 LT21

CS32 LT16

3 CS31 LT15

IS31 LT19 OR31 ALL

4 5

Friday

OR33 LT9

DB32 LT21

SC33 LT7

* This table was posted in the university web site during semester 1 2003/2004. It has been updated during Christmas vacation as follow:

Monday 9

Tuesday SC33 LT4

10

Wednesday

Thursday

Friday

CS32 LT14 SY31 LT16

AI32 LT11

AI34 PBB08

SY32 LT24

AI32 LT10

11

OR33 PBB10

IS31 LT23

12 OR31 LT11

SE31 LT15

IS31 ALL

1 SE31 LT16

AI34 PBB08 CS31 LT16

2 SY31 LT13

CS32 LT16

3 CS31 LT15

IS31 LT19

OR31 ALL

4

DB32 LT10

SC33 LT15

DB32 LT19 OR33 LT11

SY32 LT15

5

33

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix D: Level 1 Monday Tuesday Wednesday Thursday Friday

9:00-11:00 Level1-Zone1 Level1-Zone2 Level1-Zone1 Level1-Zone3 Level1-Zone2

12:00-2:00 Level1-Zone3 Level1-Zone1 Free Level1-Zone2 Level1-Zone3

3:00-5:00 Level1-Zone2 Level1-Zone3

9:00-11:00 Level2-Zone2 Level2-Zone3 Level2-Zone2 Level2-Zone1 Level2-Zone3

12:00-2:00 Level2-Zone1 Level2-Zone2 Free Level2-Zone3 Level2-Zone1

3:00-5:00 Level2-Zone3 Level2-Zone1

9:00-11:00 Level3-Zone3 Level3-Zone1 Level3-Zone3 Level3-Zone2 Level3-Zone1

12:00-2:00 Level3-Zone2 Level3-Zone3 Free Level3-Zone1 Level3-Zone2

3:00-5:00 Level3-Zone1 Level3-Zone2

9:00-11:00 Level1 Level3 Level1 Level2 Level3

12:00-2:00 Level2 Level1 Free Level3 Level2

3:00-5:00 Level3 Level2

Level1-Zone1

Level 2 Monday Tuesday Wednesday Thursday Friday

Level2-Zone2

Level 3 Monday Tuesday Wednesday Thursday Friday

Level3-Zone3

Subject allocated to Zone1 Monday Tuesday Wednesday Thursday Friday

34

Level1

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix E: All compulsory modules for the Level 2 JHSC degrees joining with School of Computing Modules (IS, Computing and Computing Science) in semester 1:

Subject

Accounting

Information System

Management studies

Module code

Module title

LUBS2200

Business Finance 2

LUSB2210

Financial Accounting 2

LUBS2220

Management Accounting 2

COMP2470

E-commerce Systems

COMP2400

Database Principles and Practice

LUBS2235

Introduction to Financial Accounting

LUBS2850

Marketing Introduction modelling for Management

LUBS2195

Studies(II)

CHEM2110

Structure and Reactivity in Inorganic Chemistry

CHEM2190

Structure and Spectroscopy

CHEM2420

Laboratory Work 2b

COMP2400

Database Principles and Practice

COMP2600

Computer Operating Systems

MATH2420

Multiple Integrals and Vector Calculus

MATH2080

Further Linear Algebra

MATH2710

Mathematical Statistics

MATH2730

Analysis of Experimental Data

Artificial Intelligence

COMP2340

Fundamentals of AI

Physics

PHYS2042

Quantum Mechanics

Computing Science

COMP2360

Theory of Computation

French

FREN2011

Language in Contexts I

LUBS2040

Theories of growth, value and distribution

LUBS2050

Industrial Economics

Chemistry

Computing Mathematics Statistics

Macroeconomic policy & performance in

LUBS2280

Britain

Economics LUBS2300

Explanation in Economics

TRAN2010

Transport Economics Mathematics and statistics for Business and

LUBS2240

Economics 2

35

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix F: All compulsory modules of BSc Honours Cognitive Science in Level 2 semester 1

Subject

Cognitive Science

Module code

Module title

COMP2340

Fundamentals of AI

PHIL2020

Meaning and Truth

PSYC2011

PSYC2040

36

Cognition, Emotion and Motivation Visual Perception and Language

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix G:

The amount of lectures and the time slot of some compulsory modules in a single teaching week: Accounting:

Artificial Intelligence:

LUBS2200 - 1 lecture, 2 hours each

COMP2340 - 2 lectures, 1 hour each

LUBS2210 - 1 lecture, 2 hours each

Information System:

LUBS2220 - 1 lecture, 2 hours each

COMP2470 - 2 lectures, 1 hour each

Management:

COMP2400 - 2 lecture, 1 hour each

LUBS2235 - 3 lecture, 1 hour each

Computing Science:

LUBS2195 - 1 lecture, 2 hours each

COMP2360 – 2 lectures, 1 hour each

LUBS2850 - 1 lecture, 1 hour each

Mathematics:

Economics:

MATH2420 - 2 lectures, 1 hour each

LUBS2040 - 1 lecture, 1 hour each

MATH2080 - 2 lectures, 1 hour each

LUBS2050 - 1 lecture, 1 hour each

French:

LUBS2280 - 2 lectures, 1 hour each

FREN2011 - 2 lectures, 1 hour each

LUBS2300 - 1 lecture, 1 hour each

Physics:

TRAN2010 - 2 lectures, 1 hour each

PHYS2042 - 1 lecture, 2 hour each

LUBS2240 - 2 lectures, 1 hour each

Psychology:

Computing:

PSYC2011 - 2 lectures, 1 hour each

COMP2400 - 2 lectures, 1 hour each

PSYC2040 - l lecture, 2 hours each

COMP2600 - 2 lectures, 1 hour each

Philosophy: PHIL2020 – 1 lecture, 1 hour each

37

Produced by: Junyan Chen

TIMETABLING MODULES FOR UNIVERSITY OF LEEDS Appendix H: Project Schedule Event

OCT

NOV

DEC

JAN

FEB

MAR

APR

MAY

Investigate current process Investigate the similar project Methodologies reading Teaching techniques search Mid project writing Interview with some department officers Evaluating tools Design Test Develop Evaluation Write up contents and draft chapter Write up project

It is expected that most of researches will be done during the first semester (before Christmas vacation). The rest of the job will be finished during semester 2. Moreover, during semester 1 it focuses on the background reading and information researches. Whereas, solving problems for the system is the keystone in semester 2.

38

Produced by: Junyan Chen