output analysis

output analysis - SYSC 4005-5001 Winter 2010 Output...

Info iconThis preview shows pages 1–7. Sign up to view the full content.

View Full Document Right Arrow Icon
1 SYSC 4005-5001, Winter 2010 Output Analysis Winter 2010 Slides are based on the texts: -Discrete Event System Simulation, by Banks et al -Discrete Event Simulation: A first Course, by Leemis and Park -Also check the book by professor Perros (free available in the course web site) especially the material on page 105.
Background image of page 1

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
2 SYSC 4005-5001, Winter 2010 Purpose ± Objective: Estimate system performance via simulation ± If θ is a measure (i.e average delay) of the system performance, the precision of the estimator can be measured by: ² The statistical error of and the width of a confidence interval (CI) for θ . ± Purpose of statistical analysis: ² To estimate the standard error or CI . ² To figure out the number of observations required to achieve desired error/CI. ± Potential issues to overcome: ² Autocorrelation, observations lack statistical independence . ² Initial conditions can affect the course of simulation ˆ ˆ
Background image of page 2
3 SYSC 4005-5001, Winter 2010 Outline ± Distinguish the two types of simulation: transient vs. steady state. ± Illustrate the inherent variability in a stochastic discrete- event simulation. ± Cover the statistical estimation of performance measures. ± Analysis of transient simulations. ± Steady-state simulations.
Background image of page 3

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
4 SYSC 4005-5001, Winter 2010 Type of Simulations ± Terminating (transient) versus non-terminating simulations (steady-state) ± Terminating simulation: ² Runs for some duration of time T E , where E is a specified event that stops the simulation. E.g. simulate the time it takes on the average till a single server queue with finite storage overflows for the first time! ² Starts at time 0 under well-specified initial conditions. ² Ends at the stopping time T E (first overflow event)
Background image of page 4
5 SYSC 4005-5001, Winter 2010 Type of Simulations ± Non-terminating simulation: ² Runs continuously, in practice over a very long period of time. ² Examples: Simulate a queue and find average delay by letting 1000000 packets to be processed (this number may be found suitable using the modelers experience…) ² Initial conditions defined by the analyst, may not have an effect to the final results! ² Runs for some analyst-specified period of time T E . ² Study the steady-state (long-run) properties of the system, properties that are not influenced by the initial conditions of the model. ± Whether a simulation is considered to be terminating or non- terminating depends on both ² The objectives of the simulation study and ² The nature of the system.
Background image of page 5

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
6 SYSC 4005-5001, Winter 2010 Stochastic Nature of Output Data ± Model output consist of one or more random variables (r v’s) because the model is an input-output transformation and the input variables are r.v.’s. ± M/G/1 queueing example: ² Poisson arrival rate = 0.1 per minute; service time ~ N( µ = 9.5, σ =1.75). ² System performance: long-run mean queue length, L Q (t) .
Background image of page 6
Image of page 7
This is the end of the preview. Sign up to access the rest of the document.

This note was uploaded on 04/16/2010 for the course SCE sysc5001 taught by Professor Lambadaris during the Spring '10 term at Carleton CA.

Page1 / 46

output analysis - SYSC 4005-5001 Winter 2010 Output...

This preview shows document pages 1 - 7. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online