Seminars

The information on this page is not completed and schedule is tentative. Please check back later for more information.

  • Title: Pursuit Evasion in Simple Polygons
    Speaker: Babar Majeed

    Date: April 7
    Time: 3:30pm - 4:30pm
    Room: CL 407

    Abstract: Pursuit evasion in simple polygons is mainly about searching for an evader in a simple polygon. Popular problems in pursuit evasion consider a room search problem and a two guard problem. Many practical applications, such as search for an evader in a dark house, rescue of a victim in a dangerous building and surveillance with autonomous mobile robots can be modeled by polygon search problems. In this seminar, first we will consider the problem of searching a mobile evader in a simple polygonal room with two doors by a boundary 2-searcher. The necessary and sufficient conditions for a simple polygonal room to be searchable by a boundary 2-searcher are presented and an algorithm for generating the motion path is proposed, if it exist. It is reported that searching a room using a boundary 2-searcher requires at most O(n2) edge traversal. Second problem which we will consider is searching a mobile evader in a simple polygon using a pair of searchers. The pair of 1-searchers has to detect the evader at least once. The necessary conditions for a simple polygonal room to be searchable by a two 1-searchers are presented.

  • Title: Simulating a Cops and Robber Game on a Street Map
    Speaker: Rahim Ali

    Date: April 4
    Time: 3:30pm - 4:30pm
    Room: CL 407

    Abstract: The cops and robber game is a perfect information game with a cop player, who controls a set of cops which are moved and placed first, and a robber player controlling a single robber and goes second, that is played on a finite connected graph. The main goal is to create a strategy for a graph that allows the cops to catch the robber while using the minimum number of cops. Using OpenStreetMap data I will present a simulation of cops and robber games on maps of various city areas with intersections acting as nodes, and show how a simulation can be used to assist in creating a minimum cop-win strategy for such a graph.

  • Title: Challenges of Card Sorting Analysis
    Speaker: Hadeel Hatin Bin Amer

    Date: March 28
    Time: 3:30pm - 4:30pm
    Room: CL 407

    Abstract: Card sorting is a technique used to collect data from participants about specific items. Nowadays card sorting could be done in a real or virtual life. There are two different types of card sorting, which are: open and closed card sorting. The main goal of card sorting is to distinguish the various approaches of cards classification. For example, card sorting can compare between the one way of thinking when people sort the cards with the different ways that are manifested in these sorting. Card sorting analysis can be done using clustering software or by looking for common patterns between the collected data. One of the challenges that most researchers face is how to analyze and interpret the card sort data.

  • Title: A Web-based Learning Support System for Rough Set Theory
    Speaker: Ying Zhou

    Date: March 24
    Time: 3:30pm - 4:30pm
    Room: CL 407

    Abstract: Web-based learning systems are gaining popular due to their convenience, ubiquity, personalization and adaptation features comparing with traditional learning environment. The subjects of most of Web-based learning systems are popular sciences. Little attention has been paid for learning cutting edge academic subjects and no such systems for learning Rough Set Theory. In this presentation, I will introduce a Web-based learning support system (named MRS) that is specifically designed for learning Rough Set Theory. I will be briefing the design principle, architecture, and implementation of MRS. A special feature, adaptation, will be emphasized. The system has the ability of adaptation to student's preference and performance by modifying the size and order of learning materials delivered to each individual. First, it tailors and sorts learning materials based on student's preference. Second, learning materials will be automatically adjusted to students of poor performance for review purposes. Additionally, it predicts estimated time of study, which will be helpful to schedule.

  • Title: Rule simplification by using structured rough set approximations
    Speaker: Mengjun Hu

    Date: March 21
    Time: 3:30pm - 4:30pm
    Room: CL 407

    Abstract: Concept analysis is a main task in rough set theory. Based on an information table that describes a set of objects with respect to a set of attributes, we can obtain two kinds of sets of objects, that is, definable sets that can be defined precisely and undefinable sets that cannot. By introducing upper and lower approximations, one can approximately define an undefinable set. Then based on approximations, several rules can be derived to indicate whether an object belongs to a given set. The internal structure of approximations decides the structure of rules. In our research, we focus on the structure information in approximations and introduce the structured approximations. Thus, rules can be simplified by using less attributes.

  • Title: Conditional Preference Networks: Representation and Inference
    Speaker: Eisa Alanazi

    Date: March 17
    Time: 3:30pm - 4:30pm
    Room: CL 407

    Abstract: Reasoning with preferences is a topic of great interest to many domains including artificial intelligence, recommender systems and decision theory to name a few. Last decade shows a great body of work in developing representation languages and efficient algorithms to reason about preferences. One notable preference formalism is the Conditional Preference Network (CPNet). CPNets are graphical models to represent complex preferences over combinatorial domains. A CPNet is a directed graph where vertices are decision variables and edges represent the preferential dependencies. This talk concerns reviewing the representation of CPNets and different inference problems related to the model.

  • Title: Route Planning Algorithms for More Efficient 3D Printing
    Speaker: Shulang Lei

    Date: March 14
    Time: 3:30pm - 4:30pm
    Room: CL 418

    Abstract: General 3D printers use FDM (fused deposition modelling) technologies due to their affordability. However, FDM extruding requires a continuous Gcode (Tool controlling code) path to avoid overhang problems. Furthermore, solid polygon slices of printable objects use infill paths to save materials while the object remains sufficiently sturdy. Applying a single extruding path whenever is possible will produce more efficient printing results and use less material with reduced overhangs. Traveling salesman solvers are generally used to construct single path from a set of points. But they trend to be slower due to higher computation complexity. In this talk, I will discuss how to use a combination of eulerian path algorithms, Chinese postman problem algorithms and traveling salesman algorithms for optimal result.



To Top of Page