Class overview How to build simulations key issues how to do quickly how to build the right model graphics virtual reality simulation are combining for some applications how to know of it s correct how to interpret output correctly component reuse DoD emphasis CS 475 575 1 Texts Park Leemis Discrete Event Simulation A First Course to be published by Prentice Hall not available in bookstores The good news is it s free CS 475 575 2 Since this is CS class focus on computational issues and how internals work focus on simulation programming languages use right tool to solve problem make new mistakes not old ones speed speed speed build models fast build fast models efficient implementations distributed parallel simulation CS 475 575 3 Central problems in simulation Build the right model what to include how much detail effects validity speed development time reuse How to get good performance basic data structure is events list issues in distributed parallel simulation How do you KNOW you can believe output bad model data wrong assumptions bad code wrong interpretation of output How do you get others to believe output advantages disadvantages of animation CS 475 575 4 Tricky stuff Bad RNGs Incorrect statistical analysis Parallel simulation CS 475 575 5 What s simulation at least in 404 Use of models to replicate behavior of real system Models can have very different objectives Scientific experimentation understand behaviors of existing system how does DNA work how did the big bang work Design a new system optimize a new system evaluate alternatives find surprises aka design flaws Enhance understanding use simplified animations Develop skills use scripted situations land a 747 with 2 engines failed CS 475 575 6 What is sim cont Merging of graphics virtual reality simulation problems are human perception speed of rendering Much scientific computing is simulation fluid flow problems understanding pollution problems in Chesapeake Bay problems are mostly numeric CS 475 575 7 Simulate to answer what if questions 1 What happens to the power grid if a tree falls on a power line in Ohio 2 What is the effect on a skyscraper if there is an earthquake 3 What is the effect of a windstorm on a bridge see google Tacoma Narrows bridge disaster 4 What happens to a tall building hit by a jet 5 what happens to a network in a heavy traffic situation 6 what happens to the traffic on the Garden state if a toll booth is eliminated 7 Is there an increase in revenues for a pizza parlor if an extra delivery person is hired 8 What happens to the federal deficit if taxes on dividends are eliminated 9 What happens to the economy if the prime rate is increased by the Federal reserve 10 what is the effect on fuel consumption if a spoiler is added to a car 11 what happens to the deer population if their natural predators are eliminated 12 How does global warming affect the migration of geese and tonnage of their droppings in NJ 13 What happens to enrollment if GE is halved CS 475 575 8 Different needs for different uses Sometimes need complex messy depictions realism important for military training learn to fly 747 Sometimes need abstract simplified models use model to enhance user s understanding extraneous details inhibit understanding Sometimes performance crucial sometimes not Sometimes correctness crucial sometimes not Some simulations are only run once some are used frequently CS 475 575 9 What is simulation in 404 Most textbooks take on simulation much too narrow Implicitly assumes simulation is used for analysis or design Ignores training education scientific experimentation recreation etc CS 475 575 10 Simulation is used to train personnel 1 pilots go into flight simulators 2 air traffic controllers 3 emergency responders CS 475 575 11 Simulation is used to entertain 1 Sim city 2 basketball hockey baseball computer games CS 475 575 12 Traditional uses of sim Simulation used when system of interest doesn t exist sim used to design system is too expensive to use for mere studies is too dangerous is too slow want to take quick look at alternatives to see which to pursue in detail long period studies say a week or a decade is too fast slow it down to see what s happening CS 475 575 13 1 2 Systems Models and Simulation cont d Ways to study a system CS 475 575 Simulation is method of last resort Maybe But with simulation there s no need or less need to look where the light is 14 1 2 Systems Models and Simulation cont d Classification of simulation models Static vs dynamic Deterministic vs stochastic Continuous vs discrete Most operational models are dynamic stochastic and discrete will be called discrete event simulation models CS 475 575 15 1 3 DISCRETE EVENT SIMULATION Discrete event simulation Modeling of a system as it evolves over time by a representation where the state variables change instantaneously at separated points in time More precisely state can change at only a countable number of points in time These points in time are when events occur Event Instantaneous occurrence that may change the state of the system Sometimes get creative about what an event is e g end of simulation make a decision about a system s operation Can in principle be done by hand but usually done on computer CS 475 575 16 Reality main consumer of sim is US Dept of Defense Software example old specify design develop validate software for Aegis cruiser basic ftns navigation communication weapons sensors situation awareness CS 475 575 17 Example radar component Design system so that at any point radar inputs to rest of system can be from a REAL radar system from history recorded sensings from previous field exercises from simulation model to experiment with different loadings situations whatever from large training exercise where inputs come from other players models people planes ships CS 475 575 18 Uses Such a systems could be used to test effect loadings from newer radar system on other ship systems evaluate correctness of other ship systems which must respond to radar train personnel by simulating dangerous expensive but carefully scripted conditions to meet training objectives evaluate human factors light loads heavy loads special circumstances CS 475 575 19 Term discrete event Previous example for most purposes can be represented at a discrete event system at t1 first blip appears at t2 pkt containing data is placed on network how to determine how long it takes after t1 for t2 at t3 pkt arrives at node i at t4 second blip