Study projects
The projects I have or are participating in at the Computer Science Department at Aalborg University.
Computer Science study at Aalborg University
I'm currently studying Computer Science at the Computer Science department (dk) at Aalborg University (dk).
During my study I have worked on different projects, concerning many different topics. I have choosen to publish my projects here on this page, since there maybe som useful information i some of the projects.
Before introducing the different projects I will apolygise that some of the projects is written in danish. Theese will be marked with a (dk). The paper- and report-descriptions will include the abstract as a description of the project.
P 1 - Automatic Scheduling (dk) 
My P1 project where about the scheduling problem e.g. planning a work schedule for all employees in a large coporate firm. The report follows the project evolution in the automated sheduling area. The project challenges som hypotheeses about problems in planning a schedule. This ends up in a algorithm that though tests is shown able to plan a schedule. We implemented a prototype of the algorithm in PHP.
Synopsis: I denne rapport følges et projektforløb som omhandler automatiseret vagtplanlægning. Projektarbejdet tager udgangspunkt en række hypoteser omkring problemstillinger vedrørende vagtplanlægning i praksis. Disse hypoteser undersøges nærmere og en algoritme til automatiseret vagtplanlægning udvikles, som et løsningsforslag til problemstillingerne. En vellykket test af vores algoritme, implementeret i PHP, blev udført. Resultatet blev en ''gyldig'' vagtplan, der tog højde for kvali kationer, personlige ønsker og løntrin for en gruppe medarbejdere, i en ktiv testcase. Hver medarbejder havde tilknyttet et antal timer, pr. uge. Afvigelsen fra det fast timetal, var i testen gennemsnitlig 10%.
P 2 - Multiple Routes Planning
My P2 project where about Multiple Routes Planning.
Abstract: Based on an interview with L.C.H. Import A/S, we have identified a route planning problem, which we regard as a Travelling Salesman Problem. The problem is represented using graph theory. In an attempt to help solve this problem, we have developed a program design which uses Christofides algorithm while creating routes in a complete graph. The program design consists among other things of a demand specification based on the identified problems, with demands which needs to be kept for the program design to be usable. The program design has been implemented in a prototype in ANSI C, which has been evaluated in regard to the demand specification. Our test results show that the prototype produces routes, less than 1.5 times longer than the optimal route, within an acceptable timespan. We can conclude that a software based solution can ease a route planning process.
Dat 1 - Scheduling Assistant
My Dat1 project where about Workload allocation, and during working with this problem, we developed a Scheduling Assistant program to ease the schedule planning.
Abstract: This project concerns a schedule planning problem at a department at Aalborg University. The result of the used object orientated analysis and development, is a software solution, Schedule- Assistant, which is a tool to help a coordinator create and maintain a schedule for a semester. The software is developed in C# and provides the possibility for direct manipulation through drag and drop of lectures in a calendar module. Furthermore the software includes a function to populate the schedule automatically. Another aspect covered during the development is serialization of cross referenced objects. During the programming phase, unit tests were used to evaluate the written code, and make sure the classes worked as intended.
Dat 2 - Sip2Sep (dk)
This Dat 2 project were about the development of a programming language and a compiler plus interperter to that language:
Synopsis: Denne rapport giver et bud på regler, der kan opstilles for at parallelisere sekventiel kode, og derved opnå implicit parallelitet. De to sprog Sip og Sep med hhv. implicit og eksplicit parallelitet, danner grundlag for en analyse af parallelitet baseret på en fuldstændig operationel semantik og en formel beskrivelse af parallelisering betragtet som et transistionssystem. Der implementeres en oversætter fra Sip til Sep. Slutteligt udføres forsøg med oversættelse og fortolkning af sekventielle programmer skrevet i Sip og deres tilsvarende parallele programmer oversat til Sep.
Dat 3 - Peer-To-Peer Podcast Distribution
My Dat3 project where about the scheduling problem e.g. planning a work schedule for all employees in a large coporate firm.
Abstract: This project concerns the development of a P2P podcast distribution system both in theory and in implementation. Concepts like replication, directory services, credit-based trading, security etc. are discussed. Based on analysis and design, a functioning P2P system, using directory services, is implemented in Java, using FreePastry as P2P middleware.
Dat 4 - Continuous Queries on Current and Predicted Locations of Moving Objects
This is my current project. A Database projects done i corporation with the database department at Computer Science at AAU
Abstract: In recent years there has been an increase in the number of advanced location based services(LBS). When the number of users becomes large, all LBS location servers suffer from hugeamounts of updates. Recently developed location update policies, like so-called velocity-vectorbasedupdate policy, in comparison to currently used methods, achieve less location updateswith the same given accuracy. There is a challenge to integrate this update technique into aLBS system and thereby significantly reduce communication traffic between clients and theserver.We present a general purpose database solution, which can be used as a core element in mostLBS systems. It employs spatial indexing on continuous queries, while using velocity-vectorbasedshared prediction update technique to provide an efficient evaluation of spatio-temporalqueries on moving objects. We present an implementation of our solution on a spatially enabledrelational database system and provide performance study results.
Dat 5 - A Data Warehouse Solution for Analysis on Indoor Tracking Data
This project concerns the design and implementation of a data warehouse for analysis on tracking data collected at Copenhagen Airports A/S. The project is done in cooperation with BLIP Systems A/S who has a Bluetooth based tracking system implemented in Copenhagen Airports A/S. First, we ntroduce
a number of technologies that can be applied for indoor tracking, then we describe how BLIP Systems A/S stores the collected tracking data and propose a data warehouse design that enables answering some business intelligence related questions provided by BLIP Systems A/S. As a part of the Extract
Transform Load, some data cleansing has been performed on the source data. Additionally, we introduce two casespeci c algorithms, FlyerFrequencyCalculation, used for calculating measures and BounceDetection, which normalizes the tracking data. By utilizing Targit BI Suite, we address some of the questions presented and provide answers through graphs. Lastly, we conclude on the project and suggests directions for further work.
Dat 6 - A Data Warehouse Solution for Flow Analysis Utilising Sequential Pattern Mining
Tracking of visitors in indoor spaces has many applications, specifically when administrating large public areas, like airports and train stations.
In this work we present a data warehouse solution, designed to store and facilitate analysis of large amounts of tracking data.
The solution is designed specifically for a data set provided by Blip Systems A/S, tracking visitors in Copenhagen airport. We propose to utilise sequential pattern mining to precompute flow information and present three different approaches, to incorporating frequent patterns for flow analysis in the data warehouse.
A prototype system is implemented to provide grounds for an extensive experimental study. The study discloses the strengths and weaknesses of the different approaches, and indicates in which cases the different approaches are applicable.
Dat 7 - eEvent - Collaborative Event Planning
This work presents an idea of collaborative event planning and develops this idea into a new venture. The idea is developed using various techniques like W5H2, Disney and especially an innovation management strategy called The Seven Circles of Innovation. The idea has been selected, since collaborative event planning between multiple peers, can be a time consuming task. When the amount of peers increase, the task of collaborative event planning can become almost impossible, due to the amount of communication needed to reach agreement.
We propose to utilise Business in Beta as the business creation strategy, and present a business plan, describing most aspects of the planned business.
A prototype is implemented and tested in a proof of concept study, involving users from the target group. The study yields that the idea of collaborative event planning could be feasible, even though there are some financial issues.