Master theses

In the list below you find a number of suggestions for master thesis topics.

Are you curious about what former master students have written their theses about?
Then follow the link to: 
Finished Master theses

 


 

Identification of ITS trends by Mining Company Summaries
CONTEXT. Researchers, companies, and public authorities meet annually at
the 18th world congress on Intelligent Transport Systems (ITS) to
combine efforts and discuss advances in the use of ITS to address key
challenges faced by transportation in general and road transport in
particular. Usually, hundreds of major ITS companies with different
solutions ranging from experimental to deployed systems participate. Due
to the wealth of proposed technologies and the variation of interests
amongst different stakeholders, it is difficult to get an overview of
the current trends.
OBJECTIVES. The main focus of each participant in the congress is
presented with a short summary. The collection of summaries are
available through the congress website or via an iPhone application. An
investigation of data mining approaches for extracting relevant
information from congress summaries to automatically identify key trends
in the ITS community will be performed. The next objective is to perform
an experimental comparison of the potentially suitable approaches.
EXPECTED OUTCOMES. The investigation is expected to result in an
analysis of data mining approaches for mining micro texts. Moreover,
experimental results will be used as a basis for identifying the most
promising approach for extracting key trends from ITS congress
summaries. An automatic key trend extraction method may, if successful,
make it simpler for the ITS community to get an overview of important
trends. In addition, the method could provide researchers with valuable
decision support when there is a need to focus research efforts.

 

Topic: Identification of ITS trends by Mining Company Summaries

Contact Person: Gideon Mbiydzenyuy

CONTEXT. Researchers, companies, and public authorities meet annually atthe 18th world congress on Intelligent Transport Systems (ITS) tocombine efforts and discuss advances in the use of ITS to address keychallenges faced by transportation in general and road transport inparticular. Usually, hundreds of major ITS companies with differentsolutions ranging from experimental to deployed systems participate. Dueto the wealth of proposed technologies and the variation of interestsamongst different stakeholders, it is difficult to get an overview ofthe current trends.
OBJECTIVES. The main focus of each participant in the congress ispresented with a short summary. The collection of summaries areavailable through the congress website or via an iPhone application. Aninvestigation of data mining approaches for extracting relevantinformation from congress summaries to automatically identify key trendsin the ITS community will be performed. The next objective is to performan experimental comparison of the potentially suitable approaches.
EXPECTED OUTCOMES. The investigation is expected to result in ananalysis of data mining approaches for mining micro texts. Moreover,experimental results will be used as a basis for identifying the mostpromising approach for extracting key trends from ITS congresssummaries. An automatic key trend extraction method may, if successful,make it simpler for the ITS community to get an overview of importanttrends. In addition, the method could provide researchers with valuabledecision support when there is a need to focus research efforts.

 


Topic: Pattern Recognition within CyberForensic TimeLab

Contact Person: Martin Boldt

Computer Forensics is mainly about investigating crime where computers have been involved. There are many tools available to aid the investigator with this task. We have created a prototype of a new type of tool called CyberForensic TimeLab where all evidence is indexed by their time variables and plotted on a timeline. We believed that this way of visualizing the evidence allows the investigators to find coherent evidence faster and more intuitively. This master thesis topic is  about enhacing this tool by adding an "intelligent" component (based on machine learning/data mining techniques) which could find interesting patterns within the distribution of the timestamps. Also, additional ways to visualize the data is also needed. To evaluate the tool some kind och user survey could be used. More information regarding the tool is available here.


Topic: File fragments analysis and identification.

Contact person: Stefan Axelsson

When doing forensics research one often ends up with a number of small blocks that previously belonged to now deleted files (i.e. by analyzing so called slack space on a hard disk). Trying to identify the type of the files (i.e. image, video, word document etc.) these blocks belong to is an active area of research in computer forensics, as a first step in trying to reconstitute the files themselves. Machine learning is often brought to bear on this problem. This project is for one to two students to try and develop or apply novel techniques in addressing the above problem.


Topic: Beslutstödsystem för störningshantering av tågtrafik

Contact person: Johanna Törnquist

Description: Tänk dig att du i Karlskrona stiger på ett tåg (f.d. Kustpilen) ner till Kastrup för att ta ett flyg till New York och du har 2 timmars marginal mellan tågets ankomst och flygets avgång. Direkt efter avgång från Karlshamn försenas tåget kraftigt av ett tillfälligt signalfel och nedsatt hastighetsgräns och tåget blir 20 minuter försenat. Som van resenär vet du att denna störning sannolikt kommer att öka i takt med att tåget närmar sig Kristianstad, där du behöver byta tåg vidare ner mot Malmö och Kastrup. Huruvida ditt tåg verkligen blir mer försenat eller har möjlighet att köra in tid (och minska förseningen) beror på vilka andra tåg ditt tåg möter samt hur tågtrafikledaren väljer att hantera förseningen genom att t ex omplanera vissa möten. En policy som ibland tillämpas är att prioritera tåg som är i tid (dvs. ditt tåg får stå åt sidan för andra tåg) men om tågtrafikledaren ser att det finns utrymme att försöka minska ditt tågs försening utan att orsaka ytterligare förseningar för andra tåg så gör tågtrafikledaren det. Hur ska dock tågtrafikledaren kunna veta vilka alternativa åtgärd som kan vara effektiva och vilka konsekvenser åtgärderna får?

Detta är ett exempel på en situation som kan uppstå på de flesta sträckor i järnvägsnätet och ibland få stora konsekvenser. Järnvägstrafik är känsligt för störningar och att hantera konsekvenserna av störningarna är väldigt komplex eftersom sambanden mellan åtgärder och effekter inte så enkla att förutse. Samtidigt är hanteringen viktig för prestandan i järnvägsnätet. En av de stora utmaningarna är således att kunna beräkna, åskådliggöra och jämföra olika lösningar t ex genom simulering av olika lösningsscenarier och grafisk/statistisk presentation av dessa.

Din uppgift är att tillsammans med forskningsprojektet OAT (Omplanering Av Tåglägen) utveckla och testa strategier för effektiv hantering av störningar i tågtrafiken. Fokus här ligger på att specificera, utforma och utveckla funktionaliteter och gränssnitt som tillsammans med existerande beräkningsalgoritmer kan användas av tågtrafikledare för att analysera olika störningsscenarier och visualisera en eller flera lösningar (dvs. en omplanerad tidtabell) i en störningssituation.

Önskvärt startar arbetet senast slutet av september (2005). Storlek på projektet kan diskuteras men lämpligtvis pågår arbetat under hela hösten 2005 och omfattar ett eller två magisterarbeten.

Eftersom en stor del av arbetat innefattar programutveckling behöver du ha god programmeringsvana i framför allt Java. Flera av de dokument du kan behöva läsa är på svenska och därför krävs det att du är svensktalande. Att du kan tillgodogöra dig engelska texter är även en förutsättning.

Projektsida för OAT
Beskrivning av magisterprojektet (pdf)


Topic: Automated Runtime Analysis of Privacy-Invasive Software
Contact person: Martin Boldt

Description: In previous work we have proposed the use of collaborative reputation systems for preventing privacy-invasive software (PIS), sometimes loosely labeled as spyware. More information about this work could be found here (http://www.softwareputation.com/). This master thesis proposal focus on runtime analysis of software to investigate whether the software holds certain characteristics e.g., if it display advertisements or if the uninstall feature is incomplete.

By monitoring software programs (e.g. downloaded from http://www.download.com/) that execute inside a virtual PC it is possible to identify characteristics about it. At the end of the analysis the software is removed by using the uninstall feature and then the system is thoroughly checked for any left components, i.e. to see if the uninstall mechanism is incorrect.

When the analysis is over a report that describe the software characteristics is constructed. This information could then be added to the reputations system to further help users. To manage this master thesis you need to have firm knowledge of both C++ and the Win32 environment. Related work could be found here (http://www.cs.washington.edu/homes/tbragin/spycrawler.pdf).


Topic: Handling parallel negotiations in production and transportation planning
Contact persons: Linda RamstedtJohan Holmgren

Description: Negotiation between agents is a crucial issue in multi-agent systems to enable coordination between the agents. The contract net protocol is a negotiation protocol which is frequently used in agent-based applications, for instance in the simulator TAPAS (see http://www.ipd.bth.se/stem/) which is implemented in JADE. However, for certain negotiation situations it seems like the contract net has some limitations, which makes implementation difficult. One such problematic negotiation situation is when two parallel negotiations are dependent of each other and must be synchronized. This negotiation situation is interesting to study since it often appears in the real world, e.g., when buying production and transportation capacity. The difficulties when trying to realize this type of negotiation may for instance appear due to how the protocol is implemented or due to flaws of the protocol.

A literature review of how this problem has been solved in previous work should be conducted by investigating protocols and platforms. The negotiation situation should be analysed and the reasons for the problems should be identified. Solutions to the problematic situations should be suggested, which for instance can imply using an existing negotiation protocol, suggesting possible ways to improve the implementation of existing protocols, or even suggest a new protocol. Finally, some of the findings should be validated in the simulator TAPAS.


Topic: State of the art in integrated production and transportation planning and the potential of integrating decision support techniques

Contact persons: Johan Holmgren

Description: It is known, and has been experienced in a project (www.ipd.bth.se/fatplan ), that it is a substantial challenge to optimize production and transportation in an integrated way. It concerns the situation where the production planning includes scheduling (lot-sizing) and the transport planning includes consideration of a fleet of vehicles (i.e. fleet management and vehicle routing). Further, since the problem includes the consideration of inventory levels, both at producer and at customers, it resembles an inventory routing problem.

A master project could be about reviewing the state of the art of operational decision support for this type of planning problems. Further to develop and test ideas of improved decision support by integration of different techniques. In particular, how can algorithmic approaches (optimization methods) be enhanced by the usage of another technique, such as, simulation, machine learning, artificial intelligence or agent technology for this type of problem?

 

Edit
Share Share