STUDENT RESEARCH PROJECTS IN SYSTEM DESIGN

Mark Sh. Levin

Instute for Information Transmission Problems, Russian Academy of Science

19 Bolshoj Karetny lane, Moscow, Russia

Keywords:

Student projects, System design, Combinatorial optimization, Decision making, Heurisctics, Application.

Abstract:

The article describes a new educational approach to support the movement from traditional educational pro-

cess to research and/or engineering (design) activity in computer science/information technology, engineer-

ing, and applied mathematics.The approach was realized as faculty course for advanced senior undergraduate

students(information technology).The basic educational ﬂow is: applied problem, mathematical model, algo-

rithm, software, computing the results, report.Twelve laboratory works support the ﬂow above.In each work

a combinatorial and/or multicriteria problem is (are) examined (e.g., multicriteria ranking, multiple choice

problem, multicriteria assignment/allocation, clustering) including applied examples, algorithm(s), Matlab

program(s).Thus students can obtain their skills in applied problems, models, and algorithms. In addition, each

student can take into account his/her inclination, motivations, and personal goals. As a result, the student can

select a part of the educational ﬂow above to prepare a modiﬁed or new version of the part(i.e., applied prob-

lem, model, algorithm).Concurrently, students obtain a skills in composition of problems/models/algorithms to

get a framework for a complex real world application. Motivated students have conducted advanced research

projects and their articles were published.

1 INTRODUCTION

The movement from traditional educational process

to research and/or design activity is often the criti-

cal part of many educational efforts in computer sci-

ence (CS) and information technology, engineering,

applied mathematics, management.

In the article our approach for the above-

mentioned movement is described. Our suggestion

is based on the course including a set of labora-

tory works in which an educational ﬂow is real-

ized: applied problem, mathematical model, algo-

rithm, software, computing the results, and report

(Figure 1). In each laboratory work a special com-

binatorial and/or multicriteria problem is examined

(e.g., multicriteria ranking, multiple choice problem,

assignment/allocation, clustering) including basic ap-

plied examples and algorithm(s) and often a basic

Matlab program (Mathworks, Inc.). As a result, each

student can understand all aspects of the problem.

The examined problems are organized as a layered

framework.

Thus students can obtain the skills in the applied

problems, models, algorithms. In addition, each stu-

dent can take into account his/her inclination, moti-

vations, and personal goals. After that the student

can select a part of the educational ﬂow above to pre-

pare a modiﬁed or new version of the part(i.e., ap-

plied problem, model, algorithm).Concurrently, stu-

dent can obtain a skills in composition of prob-

lems/models/algorithms to get a framework of com-

plex real world applications.

In is reasonable to point out the suggested ap-

proach supports problem formulation and structuring

and the skill in this domain(in our opinion)is a crucial

one.

Our material is based on the author’s course on

system design in Moscow Institute of Physics and

Technology (State University) that was implemented

in 2004...2008 (Levin, 2006b; Levin, 2007c). Gener-

ally, the course corresponds to author’s books (Levin,

1998; Levin, 2006a) and articles (Levin, 1996; Levin,

Sh. Levin M. (2009).

STUDENT RESEARCH PROJECTS IN SYSTEM DESIGN.

In Proceedings of the First International Conference on Computer Supported Education, pages 67-72

DOI: 10.5220/0001974400670072

 SciTePress

2001; Levin, 2002; Levin, 2005; Levin, 2006b; Levin,

2007a; Levin and Danieli, 2005; Levin and Firer,

2005; Levin and Last, 2006; Levin and Nisnevich,

2001; Levin and Sokolova, 2004). In recent years, re-

sults of student research projects have been published

as articles (Levin and Khodakovskii, 2007; Levin and

Leus, 2007; Levin and Merzlyakov, 2008; Levin and

Safonov, 2006; Levin and Vishnitskiy, 2007).

Figure 1: Educational ﬂow-chart.

Applied problem(s)

(case studies, real world

applications)

Entry 1

(inclination for

engineering,

CS, management)

Mathematical model(s)

(or frame of models)

Entry 2

(inclination for

mathematics)

Solving scheme:

algorithm(s), procedure(s)

Entry 3

(inclination for

algorithms)

Software

Entry 4

(inclination for

software

development)

Report (a basis for articles,

presentations)

Clearly, basic books in several corresponding do-

mains are used as well, for example:(i) combina-

torial optimization (Cela, 1998; Garey and John-

son, 1979; Kelleler and Pisinger, 2004; Martello and

Toth, 1900);(ii) multicriteria decision making (Fish-

burn, 1970; Keeny and Raiffa, 1976; Roy, 1996;

Saaty, 1998; Steuer, 1986);(iii) design frameworks

(Altshuller, 1984; Ayres, 1998; Jones, 1981; Zwicky,

1969);(iv) systems engineering (Sage, 1979).

The course was realized for very good educated

and motivated students, but the approach may be used

(as a simpliﬁed version)for other students groups as

well. Note the course materials and results (including

a set of published student research articles) are acces-

sible via Internet.

2 COURSE ON SYSTEM DESIGN

The course consists of the following parts:(i) lec-

tures,(ii) laboratory works(here good results can be

prepared as papers for conferences and/or journals),

and (iii) examination. Lecture materials correspond

to the following:

1. Issues of systems engineering, system life cy-

cle.

2. System analysis and multicriteria decision

making including principles of system analysis, de-

cision making framework, methods for multicriteria

choice/ranking (utility function approach, methods

based on revelation of Pareto-effective solutions, out-

ranking techniques).

3. Optimization and combinatorial optimization

(including basic models as follows: knapsack prob-

lem, multiple choice problem, scheduling, satisﬁabil-

ity problem, graph coloring, clique problem, covering

problem, clustering, assignment/allocation problem,

travelling salesman problem (TSP), design of span-

ning structures).

4. Basic design frameworks and approaches:(a)

optimization approaches, multidisciplinary optimiza-

tion, parameter space investigation method, morpho-

logical analysis and its modiﬁcations, evolution ap-

proaches;(b) frameworks for system improvement;

and (c) modeling of system evolution/development.

5. Additional system issues: requirements en-

gineering, system evaluation and diagnosis, mainte-

nance,testing.

Course environment (computer class and Internet

site)involves the following parts:(i) lecture notes(at

course site and in computer class);(ii) program sup-

port:(a) basic software (editors, etc.),(b) network en-

vironment (Internet, etc.), and (c) Matlab environ-

ment(site of Mathworks, Inc.); and(iii) additional ma-

terials:(a) research articles;(b) support materials as

examples:to design student homepages, to prepare re-

ports on laboratory works, to prepare student articles

(a set of student published articles).

Twelve laboratory works are included into the

course:

Laboratory work 1.

Part 1A: Introductory work (computer environ-

ment, design of student homepages, preparation of

presentations).

Part 1B: Hierarchical morphological design of a

modular system (Levin, 1996; Levin, 1998; Levin,

2001; Levin, 2002; Levin, 2005; Levin, 2006a;

Levin, 2008; Levin and Firer, 2005; Levin and Kho-

dakovskii, 2007; Levin and Last, 2006; Levin and

Nisnevich, 2001; Levin and Sokolova, 2004; Levin

and Vishnitskiy, 2007; Ritchey, 2006).

Laboratory work 2. Multicriteria ranking (utility

function approach, Pareto-based approach, outrank-

ing technique as ELECTRE).

Laboratory work 3. Multicriteria knapsack prob-

lem.

Laboratory work 4. Method of proximity to ideal

point(s).

CSEDU 2009 - International Conference on Computer Supported Education

Laboratory work 5. Hierarchical clustering (Jain,

1999; Levin, 2007b).

Laboratory work 6. Multicriteria multiple choice

problem (Levin and Safonov, 2006; Poladian, 2006).

Laboratory work 7. Hierarchical ordinal eval-

uation of composite system (hierarchical method

based on integration tables) (Levin, 2006a; Levin and

Danieli, 2005).

Laboratory work 8. Composite applied example:

clustering and multiple choice problem.

Laboratory work 9. Assignment/allocation prob-

lem.

Laboratory work 10. Composite applied exam-

ple: clustering & clustering & allocation & multiple

choice problem (Levin and Merzlyakov, 2008).

Laboratory work 11. Travelling salesman prob-

lem.

Laboratory work 12. Additional work on the

topic selected by him/herself (e.g., heuristics, ge-

netic algorithms, method based on space ﬁlling

curves, cross-entropy method, quadratic assignment

problem, covering problems, real world applications).

In each laboratory work, students have to do the

following:

(1) to understand the material: applied problem,

model, solving scheme (algorithm, framework),

(2) to develop software and to test it,

(3) to prepare a numerical example (or a real world

application),

(4) to compute the results, and

(5) to prepare the report.

For some basic laboratory works there are some

basic Matlab programs, for example: multicri-

teria ranking (utility function method, revelation

of Pareto-effective solutions, outranking techniques,

e.g., method Electre), heuristic for knapsack prob-

lem, heuristic for multiple choice problem, hierarchi-

cal clustering (agglomerative algorithm). As a result,

each student has to improve the basic Matlab program

to get his/her resultant program. For more compli-

cated (i.e., composite) laboratory works (for example:

1, 6, 8, 10) each student can combine his/her Mat-

lab programs to get the resultant composite software.

Note students hav their opportunity to choose a soft-

ware environment (e.g., C, Java).

Our framework consists of the following layers:

Layer 1. Basic combinatorial problems and multi-

criteria decision making problems: multicriteria rank-

ing, multicriteria knapsack problem, ’ideal point’

method, and system evaluation/diagnosis.

Layer 2. Advanced models (e.g., clustering, mul-

ticriteria multiple choice problem, multicriteria as-

signment/allocation, TSP)

Layer 3. Composite models/procedures (as ba-

sic solving composite scheme consisting of prob-

lems/models): (i) Hierarchical Morphological Mul-

ticriteria Design (HMMD)(ranking, combinatorial

synthesis) to design modular systems, (ii) design

of a modular solving strategy (e.g., a partition-

ing/synthesis heuristic for problem), (iii) system up-

grade, (iv) multistage design (two-level HMMD),

(v) system evolution/forecasting, and (vi) special

multistage composite framework (clustering, assign-

ment/location, multiple choice problem), etc.

Figure 2 illustrates the corresponding framework

over laboratory works.

Figure 2: Framework over laboratory works.

Work



*



:

Work



*

Work



*



:

Work

Site of the course involves the following (http:

//www.mslevin.iitp.ru/SYSD.HTM):(i) description of

the course, (ii) lecture notes, (iii) students lists

with corresponding information on the students’s re-

sults/achievements, (iv) references to student home-

pages (in the case of their existence),(v) references to

signiﬁcant bibliography and Web-sites, and (vi) ref-

erences to student published articles (as examples for

future research works).

3 RESEARCH PROJECTS

Composite laboratory works are used as a funda-

mental for student research projects. Generally,

there are four possible directions for the student

research: (a) real world application (an engineer-

ing/CS/management research), (b) new or modiﬁed

model (research in mathematics or computer sci-

ence), (c) new or modiﬁed algorithm/procedure(solv-

ing scheme) (research in mathematics or computer

science), and (d) new or modiﬁed program (research

in applied computer science).

STUDENT RESEARCH PROJECTS IN SYSTEM DESIGN

As a result, each student can choose the certain kind

of his/her research. This choice process is based on

student inclination and/or experience. In the simplest

situation, students can use a simple numerical ver-

sion of an applied problem (e.g., multicritertia rank-

ing, knapsack, multiple choice problem, assignment,

TSP).

On the other hand, each student can examine a

special real world application. This step is crucial

for composite problems in the following laboratory

works: 1, 5, 6, 7, 9, 10, 11, 12. The example list of

student research projects involves the following: (1)

sport applications (e.g., organization of sport events,

planning of sport activity), (2) educational application

(computer class, interactive educational software),(3)

software development (e.g., software for modeling

of signals), (4) hardware applications (e.g., com-

puter memory), (5) communication systems (analy-

sis of communication protocols, routing, devices for

computer networks, connection of clients to network,

improvement/upgrade or extension of a network),(6)

sensors and wireless systems (telemetry system, wire-

less sensor), (7) VLSI design (planning of test pro-

cesses, planning of manufacturing), (8) private elec-

tronic devices (digital camera, notebook, car, mobile

phone), and (9) Web-based application (information

system, Internet-based station), etc.

In the case of student inclination for algorith-

mic studies, it is possible to design new or modiﬁed

algorithms (e.g., student works: for clustering, for

TSP, for covering problems, for multicriteria multi-

ple choice problem, for quadratic assignment prob-

lem, for graph coloring).

In the ﬁeld of algorithm studies, important and

useful research projects correspond to comparison of

various algorithms for the same problem (e.g., for

multicriteria multiple choice problem, for multicrite-

ria assignment problem, for multicriteria Steiner tree

problem).

It is reasonable to point out in laboratory work

12 some topics of students research projects are tar-

geted to algorithms, for example: algorithms for rout-

ing in communication networks; experimental com-

parison of heuristics for TSP (or quadratic assign-

ment problem): genetic algorithm, ant colony opti-

mization, cross entropy method; algorithms based on

space-ﬁlling curve; and algorithms for scheduling in

wireless sensor networks. In this case, the student has

to study corresponding materials (articles, books, In-

ternet) and to conduct the algorithmic research with-

out assistance (on his//her own).

Thus the course contains several entries for stu-

dent research projects (Figure 1):

Entry 1 (Application): new real world application

(on the basis of existing problem(s), model(s), algo-

rithm(s), software).

Entry 2 (Mathematical Models):model(s) (new or

modiﬁed mathematical model).

Entry 3 (Algorithms): solving scheme and/or their

analysis (theoretical, experimental).

Entry 4 (Software): program.

Figure 3 illustrates basic and advanced student re-

search project ﬂows.

Figure 3: Basic and advanced research ﬂows.

Advanced

student

research

ﬂow

Basic

student

ﬂow

666

Basic

applied

examples

(from

course)

Basic model

(or model

framework)

(from

course)

Basic

algorithm/

procedure

(from

course)

New

student

applied

research

example

Basic model

(or new one:

multicriteria,

uncertainty,

dynamics)

Basic

solving

procedure

(or new

one)

Figure 4 depicts typical domains for new student re-

search applied problems where students have their ex-

perience.

Figure 4: Domains for new applied problems.

Sport

Art (music,

dance, etc.)

Education











Communication Software

Electronics

Information

systems

Urban

environment

Note some advanced student projects have been pub-

lished as articles (Levin and Fimin, 2007; Levin and

Khodakovskii, 2007; Levin and Leus, 2007; Levin

and Safonov, 2006; Levin and Vishnitskiy, 2007;

Levin and Ryabov, 2007).

4 DISCUSSION

Generally, there exist four basic levels of educational

process (Nilsson, 1971): (1) by instructions, (2) by

CSEDU 2009 - International Conference on Computer Supported Education

explanation, (3) by examples, and (4) by creation. Ta-

ble 1 depicts correspondence of support educational

materials to levels above in our course.

Table 1: Educational levels and course support.

Educational

level

Support processes and

materials

1.by instructions

(a) lectures,

(b) books, articles,

2.by explanations

(a) lectures,

(b) laboratory works,

other students)

3.by examples

(a) lectures,

(b) laboratory works,

lectures, articles, student

works, Internet, etc.)

4.by creation

materials for educational

levels 1, 2, 3

Note a library of examples-analogues is a crucial sup-

port material (including previous student projects and

corresponding published student articles).

Evidently, the 4th level above is the most impor-

tant for student research projects, but the usage of this

level is not easy for majority of students.

Our course involves three basic parts: (a) ap-

plied problem, (b) mathematical model or frame of

the models, and (c) algorithm(s) and/or solving pro-

cedure. The levels of above-mentioned educational

process can be targeted to a certain basic part of the

course. Clearly, the examination of an applied prob-

lem is often more easy for students and this approach

is the basic one in the course. Thus, students can ana-

lyze their experience and select a certain applied prob-

lem to use basic model(s) and basic algorithm(s) (or

solving procedure) from the course. This way is very

useful for engineering, management, and CS students

because it can be based on obtained courses and skills.

On the other hand, a small part of students (about

4...10 percent) can have their inclination and motiva-

tion to mathematical models and/or algorithms. In

this case they can be oriented to creation at the level(s)

of models and/or algorithms.

In addition, it is reasonable to point out a prospec-

tive technique for organization of educational envi-

ronment (a system professor-student): an educational

process is based on a level i (i = 1, 2, 3), but a student

is thinking that the educational process corresponds

to more higher level, i.e., > i.

This technique is often crucial to increase the level of

student creativity (i.e., creativity level of student re-

search projects). As a result, the student will not be

afraid to create.

Finally, it is necessary to point out student mo-

tivation is the main and crucial basis for successful

student research projects.

5 CONCLUSIONS

In the paper, our approach to student research projects

has been described. Laboratory works are used as

a basis for several kinds of research projects (e.g.,

real-world applications, mathematical models, algo-

rithms). As a result, each student can choose the cor-

responding kind of research or a combination of the

research kinds. Evidently, this student choice is based

on his/her inclination and/or experience and/or moti-

vation. The educational approach above can be used

for organization of team research projects as well.

It is reasonable to point out the suggested educa-

tional approach is very useful for good educated and

motivated students for movement from learning stage

to research/design stage.

The future research directions are the following:

(1) extension of the set of basic models by other

combinatorial optimization problems (e.g., covering

problem, the longest path problem) which correspond

to design problems at all stages of system life cycle

(including requirements engineering, system testing,

maintenance, recycling);

(2) usage of problems under uncertainty (e.g.,

probabilistic and/or fuzzy estimates);

(3) new real world applications;

(4) usage of AI techniques;

(5) further extension of the library of student re-

search articles; and

(6) preparation of a new course for graduate stu-

dents with extended set of basic problems/models

’Combinatorial problems in system conﬁguration de-

sign’ (Levin, 2009).

REFERENCES

Altshuller, G. S. (1984). Creativity as an Exact Science.

Gordon&Breach, London.

Ayres, R. Y. (1998). Technological Forecasting and Long-

Time Planning. McGraw-Hill, New York.

Cela, E. (1998). The Quadratic Assignment Problem.

Kluwer, Dordrecht.

Fishburn, P. C. (1970). Utility Theory for Decision Making.

Wiley, New York.

STUDENT RESEARCH PROJECTS IN SYSTEM DESIGN

Garey, M. R. and Johnson, D. S. (1979). Computers

and Intractability. The Guide to the Theory of NP-

Completeness. W.H.Freeman and Company, San

Francisco.

Jain, A. K. (1999). Data clustering: a review. ACM Com-

puting Surveys, 31(3):264–323.

Jones, J. C. (1981). Design Methods. Wiley, New York.

Keeny, R. L. and Raiffa, H. (1976). Decisions with Multiple

Objectives: Preferences and Value Tradeoffs. Wiley,

New York.

Kelleler, H. and Pisinger, D. (2004). Knapsack Problem.

Springer, Berlin.

Levin, M. S. (1996). Hierarchical morphological multicri-

teria design of decomposable systems. Concurrent

Eng.: Res. and Appl., 4(2):111–117.

Levin, M. S. (1998). Combinatorial Engineering of Decom-

posable Systems. Kluwer, Dordrecht.

Levin, M. S. (2001). System synthesis with morphological

clique problem: fusion of subsystem evaluation deci-

sions. Information Fusion, 2(3):225–237.

Levin, M. S. (2002). Towards combinatorial planning of

human-computer systems. Appl. Intell., 16(3):235–

247.

Levin, M. S. (2005). Modular system synthesis: example

for composite packaged software. IEEE Trans. on

SMC - Part C, 35(4):544–553.

Levin, M. S. (2006a). Composite Systems Decisions.

Springer, New York.

Levin, M. S. (2006b). Course ’system design: structural

approach. In 18th Int. Conf. Design Methodology and

Theory DTM2006, DETC2006-99547.

Levin, M. S. (2007a). Combinatorial technological systems

models (examples for communication system). In 2nd

Int. Conf. ICSEM’07.

Levin, M. S. (2007b). Towards hierarchical clustering. In

CSR 2007, LNCS 4649. Springer.

Levin, M. S. (2007c). Towards teaching of system de-

sign. Inform. Technol.&Comput. Syst., 2:89–94 (in

Russian).

Levin, M. S. (2008). Morphological approach to electronic

shopping. In IEEE Reg. 8 Int. Conf. Sibircon-2008.

IEEE.

Levin, M. S. (2009). Combinatorial problems in system

conﬁguration design. Autom. and Remote Control,

70(3).

Levin, M. S. and Danieli, M. A. (2005). Hierarchical de-

cision making framework for evaluation and improve-

ment of composite systems. Informatica, 16(2):213–

240.

Levin, M. S. and Fimin, A. V. (2007). Hierarchical design

of ﬁre alarm wireless sensor element. In 7th Int. Conf.

CAD/CAM/PDM-2007. Inst. of Control Problems.

Levin, M. S. and Firer, M. (2005). Hierarchical morpholog-

ical design of immunoassay technology. Computers in

Biology & Medicine, 35(3):229–245.

Levin, M. S. and Khodakovskii, I. A. (2007). Structural

design of the telemetry system. Autom. and Remote

Control, 68(9):1654–1661.

Levin, M. S. and Last, M. (2006). Design of test inputs

and their sequences in multi-function system testing.

Appl. Intell., 25(1):105–124.

Levin, M. S. and Leus, A. V. (2007). Modular design of

integrated security system. In Int. Workshop DCCN

2007, volume 2. Inst. for Inform. Transmission Prob-

lems.

Levin, M. S. and Merzlyakov, A. O. (2008). Composite

combinatorial scheme of test planning (example for

microprocessor systems). In IEEE Reg. 8 Int. Conf.

Sibircon-2008. IEEE.

Levin, M. S. and Nisnevich, M. L. (2001). Combinato-

rial scheme for management of life cycle: example for

concrete macrotechnology. Intell. Manuf., 12(4):393–

401.

Levin, M. S. and Ryabov, V. A. (2007). Morphological

tree model for communication protocol (example). In

7th Int. Conf. CAD/CAM/PDM-2007. Inst. of Control

Problems.

Levin, M. S. and Safonov, A. (2006). Design and redesign

of conﬁguration for facility in communication net-

work. Inform. Technol.&Comput. Syst., 4:63–73 (in

Russian).

Levin, M. S. and Sokolova, L. (2004). Modular system

synthesis: example for composite packaged software.

Computer Meth. & Progr. in Biomedicine, 73(1):3–11.

Levin, M. S. and Vishnitskiy, R. O. (2007). Towards mor-

phological design of gsm network. Inform. Processes,

7(2):182–189.

Martello, S. and Toth, P. (1900). Knapsack Problem: Al-

gorithms and Computer Implementation. Wiley, San

New York.

Nilsson, N. J. (1971). Problem Solving Methods in Artiﬁcial

Intelligence. McGraw-Hill, New York.

Poladian, V. (2006). Task-based adaptation for ubiquitous

computing. IEEE Trans. on SMC - Part C, 36(3):328–

340.

Ritchey, T. (2006). Problem structuring using computer-

aided morphological analysis. J. of the ORS,

57(7):792–801.

Roy, B. (1996). Multicriteria Methodology for Decision

Aiding. KLuwer, Dordrecht.

Saaty, T. L. (1998). The Analytic Hierarchy Process.

McGraw-Hill, New York.

Sage, A. P., editor (1979). Systems Engineering: Method-

ology and Applications. IEEE Press, New York.

Steuer, R. L. (1986). Multiple Criteria Optimization: The-

ory, Computation, and Application. WIley, New York.

Zwicky, F. (1969). Discovery Invention, Research Through

the Morphological Approach. McMillan, New York.

CSEDU 2009 - International Conference on Computer Supported Education