COMP763 Reading Presentation Notes Yuan Jin Presentation 1 Statecharts modelling of TankWars Silvia Mur Blanch Model characters (tanks) using class diagram Model detailed bahavior using state chart formalism Model the components of tanks and assemble Class relations event sequences Time slicing vs Event Driven Time slicing vs Continuous (model freedom) Choose the level of abstraction Events; layers; data (filtering) Choose a formalism 1) State/event-based 2) Avialability of simulators 3) Well-known 4) Modularity 5) Antomanous/reactive behvior 6) Notion of itme. Presentation 2 Hybrid (Differential Equations - Statecharts) modelling in Anylogic Alexandre Denault 1. AnyLogic AnyLogic is a tools brings together System Dynamics, Process-centric and Agent=Based approaches within one modeling language and one model developement environment. Different application needs different level of modeling 2. Agent-Based Modeling Decentralised and individual centric (Pros and Cons: simulation is costly) 3. Discrete Event Modeling AnyLogic implement this by Statechart; Timer; Plain variables. 4. Continuous Time Modeling Sonething is not suitable cing discrete system (like ¡°swing¡±) Use various variables and equations 5. Hybrid Discrete/Continuous Modeling AnyLogic all back to Java Equation System: solving equations (certain threshold. Diceret)<-Hybrid model 6. Wandering Elephant Example: Tennis Ball in the water Presentation 3 Solving Static Sets of Non-Causal Equations Andrew Casey 1. What is non-causal equations It is solving 1) Simply 2)Canonicalize 3)Assign Causality 4)Order Equations 5)solve 2. Simplify is to 1)Move everything to the LHS 2)Re-express with fewer operators 3)Apply simplification rules 4)Fold constants 5)Distribute constants 6)Gather like terms Issues: 1)There's no simplest form 2)Some simplications will not be performed 3)Two expressions may be equal and yet not have the same canonical form 3. Canonicalize: 1)Constants 2)Names 3)Sums 4)Products 5)Powers 4. Causality Assignment: Build a flow graph Assign Causality: Find a pth from source to sink, reverse the edges and repeat 5. Order equations: 1)Build a dependency graph 2)Topologically sort the strongly-connected componenents There're also some corresponding issues. 6. Dynamic Sets: 1)To make the system more responsive by reducing sth 2)To update causality assignment and equatio norder dynamically 3)To solve equations in parallel 4)To iterate a constant number of times instead of until fixed point. Presentation 4 Testing of Model Transformations Amr Al Mallah OO vs MDA Proble/Domain Domain UML DS-L Lanaguage UML Running Program Running Program How to Transform LHS -> RHS graphic or DEVS MOF ¨C MetaModel Meta-object facility Testing transformation Aim: error & Confidence How?: white box (MTP) & black box MetaModel coverage: Coverage item Systematic approach Bacterio logical approach Challenge Generation of test data automatically Compring the expectedmodel to actual model(model difference, semantic difference) Presentation 5 Modelling of Aspect Weaving Wisam Al Abed Aspect Oriented Modeling Model ¡®concerns¡¯ so that they are modular and reusable Aspect model: Linking Weaving: make ¡­ classify ¡­ tracable When and how to weave? Base aspect (target) Bind/instantiate bind temporary parameter to model element thus create a context-specific aspect model. Different configuration model Temporary weaved (parameters) <- used in reusable aspect model as generic place holder the value need in target model. Inter-Aspect dependency declaration Copy and put into tracable .. ASPECTS ARE REUSABLE Template parameter preserving aspect-aspect weaving Weave to 2 aspect together Late dependency approaches Presentation 6 Heterogeneous Modelling and Simulation in Ptolemy Yanwar Asrigo 1. Heterogeneous Modeling and Simulation Example: Embedded system Each model typically represent only one aspect of the entire system How to evaluate the behavior of the whole system? 2. Ptolemy Approach Hierarchically Heterogeneous a)Divide a complex model into a tree of nested submodels b)Each level is composed of a network of interacting components 3)Network at each level is locally homogeneous, but allowing defferent interaction mechanisms to be specified at different levels in the hierarchy 3. Ptolemy Design Actor-oriented model and design Actor execution Domains or model of computation Actor polymorphism Presentation 7 Procedural Modelling of City Generation Riry Pheng 1. Context 1). Problem Statement Desirem for more: detail; realism; scale. 2). Initial Solution Hire more artists but generates some problems 3). Proposed Solution Use Procedural techniques a) describe entity in terms of instructions b) call instructions when needed 2. Types of Procedural Techniques 1). Fractals 2). L-Systems 3). Perlin Noise 4). Tiling 5). Voronoi Texture Basis 3. How to Apply in City Generation 1). Grid layout 2). L-Systems 3). Template-based 4). Split grammar 4. Wrap-up Benefits and project focus Presentation 8 Interconnection Network modelling and simulation Sina Meraji 1. Introduction 1) Interconnection networks A set of processors with local memories which communicate through a network 2) Topology: Direct and indirect 3) Switching: Store and forward (Packet switching) and Wormhole switching 4) Routing a) Topologies are simple and regular b) Very low delay and extra high bandwidth are needed c) Simple routing algos are developed d) Example: e-cube routing for hypercube networks 2. Xmulator 1) Multilayer architecture: 4 layers 2) Listener-based intergration: XML+simulator and Simulator engine 3. Conclusions and future work 1) Design a UI for Xmulator using AToM3 2) Using model transformation to generate the input XML file for Xmulator 3) Developing some routing protocols for interconnections networks 4) User can change input parameters to see their effects on performance Presentation 9 The Personal Universal Controller Yuan Jin Omitted because this is done by myself. Presentation 10 Domain-specific modelling with MetaEdit+ Willer Travassos Motivation: To slove problems like, 1. Lack of model integrity ->consistency check 2. Lack of multi user access 3. Lack of varied object representation 4. Lack of reportary technique Meta Edit+ is a true meta-model environment 5 different categories of tools Design Principle: Object Oriented Modeling, (class, relation) Data Independence Object Repository Architecture: Meta Engine Easier and integration Duplication Database -> save models in the database GOPRRR Language Presentation 11 Pounamu Chunhui Han 1. Introduction to Pounamu 1) Pounamu is a meta-tool for realizing domain-specific visual language (DSVL) environment. 2) Pounamu is implemented in Java, using the Xerces XML parsing libraries, Swing user interface packages and Java web services toolkit. 2. Pounamu structure 1) Shape designer 2) View designer 3) Meta-model designer 4) Event handler designer 3. Pounamu features 1) Zoomable view 2) Web service-based API and Thin client web-based editing interface 3) Collaborative editing of models Presentation 12 VMTS Xiaoxi Dong Presentation 13 Rhapsody Statecharts Raphael Mannadiar 1. The Basics Classes, States, Transistions, Actions and Steps 2. Basic System Reaction A simple step, Events, Triggered Operations, 3. Compound Transitions AND, OR connectors, Defaults 4. History 5. Transition Scope 6. Conflicting Transitions 7. Basic Step Algorithm 8. Multi-Threaded System Presentation 14 Model Differences Ulf Schwekendiek Presentation 15 Kermeta Jun Li Presentation 16 AI modelling Kyle Li 1. Introducing game customization 1). The definition for a Video Game and NPC 2). Current approach: a) more related to the automatic adjustment of the difficulty level of action or strategy games b) focus on RPGs. 3). Customizing NPCs: appearance, abilities, animation, interaction and environment 2. Some concepts 1) What is temperament 2) David Keirsey's theory 3) Reason about temperament 4) Ontologies 3. The model 1) Automatic NPC customization model 2) The customization process 4. Conclusions Limitations and Extensions Presentation 17 The Queueing Petri Nets formalism Kamal Zellag